Skip to main content

Full text of "The Wright Flyer : an engineering perspective"

See other formats

(/) ^ u * "*" Vt t^t Ui 4— 


CO . . 2 \ CO _ CJ CO 


^W%> a: fig *K* 

v m^ «f Isfe: a°i 


Z _■ Z 

inniSNi nvinoshiiiais saiavaan libraries Smithsonian institution NoiiniusNi nvinoshiiws s: 

r- z r; Z r Z r z 

— ,*C<< r> srZsvZr^ _ ^<vTso^\ O /SS»vS\ 




r,/ Jn^y m n^ co ^n^vy rn Xg\ pc^x ~ "'3^ m ^ru^ — X^) ja ^X m 

CO \ ± CO _ CO ± CO 





IlifUllSNI NVINOSHillMS S3iyvya 
— co 

CO ^^ 



5 > 


co ^ co — co 

O — ^£l£^ O 

Z * _J Z _l 2 -J Z 


— — 2: ^ ^ co — co £: co 


CO Z \ CO Z 

5 4& - 

co — Z (/) " Z co ••** Z co 


■ 1 


5 ^—-^ _, Z _i z ^— ^ _j z 

aiiniiisNi nvinoshiii/ms saiavaan li brar i es_smithsonian_institution NoiiniusNi NviNOSHiiws s: 





0iiniusNrNviN0SHiiiAjs u 's3 i y v a a n" l i b rar i es " "Smithsonian _ institution NoiiniusNi nvinoshiiws s 

co ~ ^ z — *" z 

Lo -v<^\. ; uj ^^r^ 2 X^^X u ...... co /c£2£&\ ^ /^H^\ co 

^zz±a^' "*?!■ *— ' ^~~_^-; — — ^dLii-^ CD 


Z r- Z 



z m \^^/ z ^ ,-v. Z ^ Z w 

0iiniiiSNi~NviN0SHiiiAjs s 3 i y vy a n ~ L ' B R AR ' Es Smithsonian institution NoiiniusNi nvinoshiiws^s: 



^ Z \ OT 2 °° 


r- — r- z 


00 — co \ 5 — co — ± w 

NouniiiSNi nvinoshiiws S3iavaan libraries Smithsonian institution NouniiiSNi nvinoshiiws 

z co z . ... co _ co 2: cr 

c 2 go * z co ' *■• Lo 




Noij.nii_SNi_NviNOSHiiws saiavuan libraries Smithsonian institution NouniiiSNi nvinoshiiiais' 

co ± oo £ oo _ oo ^ Z 


„ _ Z « CO Z CO z CO _, 

oo l<?f 




\ z /■« ^ 



T CO 2 CO 

_ /&- 

o /? 

| ^ 


z ^— r— ^ °° 5 v co — oo 

vioiiniusNi^NviNOSHiiws S3iavaan libraries Smithsonian institution NoiiniusNi" J NViNOSHii 

^ ^<rrrr\ ? \ w 2 coz, co 7 


*>.. ^ 







An Engineering Perspective 

National Air and Space Museum 

Published for the 

National Air and Space Museum 

by the 

Smithsonian Institution Press 

Washington, D.C. London 




An Engineering Perspective 

Edited by 

Special Advisor for Technology, 
Department of Aeronautics 
National Air and Space Museum 

With Contributions by 



The Wright Cfilyer 

An Engineering Perspective 

Published for the 

National Air and Space Museum 

by the 

Smithsonian Institution Press 

Washington, D.C. London 



Edited by 

Special Advisor for Technology, 
Department of Aeronautics 
National Air and Space Museum 

With Contributions by 

Library of Congress Cataloging-in-Publication Data 

Main entry under title: 

The Wright Flyer: an engineering perspective. 

Proceedings of a symposium held at the National 
Air and Space Museum, Dec. 16, 1983. 

Contents: The Wright Brothers / John D. Anderson, Jr. 
— Aerodynamics, stability, and control of the 1903 
Wright Flyer / F.E.C. Culick and H. Jex — Longitudinal 
dynamics of the Wright Brothers early Flyers / 
Frederick J. Hooven — [etc.] 

Supt. of Docs, no.: SI 9.2:W93/2 

1. Airplanes — Design and construction — Congresses. 
2. Wright Flyer (Airplane) — Congresses. I. Wolko, 
Howard S. II. Anderson, John David. III. National Air 
and Space Museum. 

TL671.2.W75 1985 629.133'343 85-600290 

ISBN 0-87474-979-4 (pbk.) 

© 1987 Smithsonian Institution. All rights reserved. 

Edited by Jeanne M. Sexton 

The paper in this book meets the guidelines for permanence 
and durability of the Committee on Production Guidelines for 
Book Longevity of the Countil on Library Resources. 

Unless otherwise noted, information about illustrations can be 
obtained from the National Air and Space Museum, Smithsonian 

Cover: Three-view drawing of the Wright Flyer. (National Air 
and Space Museum; gift of the Wright Estate) 


Acknowledgments vii 

Foreword j x 


The Wright Brothers: The First True Aeronautical Engineers 1 


Aerodynamics, Stability, and Control of the 1903 Wright Flyer 19 


Longitudinal Dynamics of the Wright Brothers' Early Flyers: 

A Study in Computer Simulation of Flight 45 


Propulsion Systems of the Wright Brothers 79 


Structural Design of the 1903 Wright Flyer 97 



The publication of any edited volume depends upon 
the coopetative efforts of many individuals. The 
following were of great assistance during preparation 
of this work. Walter J. Boyne, Director of the National 
Air and Space Museum, authorized the symposium 
and encouraged the editor at every stage. E.T. "Tim" 
Wooldridge, Chairman, Department of Aeronautics 
of the Museum, provided constant and unfailing 
cooperation. Claudia Oakes, Associate Curator of 
Aeronautics, ably attended to all administrative mat- 
ters. Dr. Tom D. Crouch, Curator of Aeronautics, 
offered many helpful suggestions. Helen C. McMahon 
was instrumental in arranging for publication of the 
work. Special thanks are due Susan L. Owen, who 
assisted with manuscript preparation. 




The National Air and Space Museum (NASM) is 
privileged to be charged with the responsibility to 
collect, preserve, restore, and exhibit the finest as- 
sembly of historic aircraft in the world. At present the 
collection contains about 300 full-size airplanes. These 
range from a Lilienthal hang glider of 1894 to the 
Dassault Falcon 20 fanjet, and include some of the 
most famous aircraft in the history of flight: The Spirit 
of St. Louis, the Winnie Mae, the Enola Gay, the Bell 
X-l, the Curtiss NC-4, the Vin Fiz, Amelia Earhart's 
Vega, Charles and Ann Lindbergh's Tingmissartoq — 
the list goes on. 

To single out from this distinguished collection one 
machine as our most precious possession might be 
thought impossible. This is not the case. When 
exhibits for the present NASM building were being 
selected in the mid-1970s, there was never any doubt 
as to which airplane would hang as the centerpiece 
of the entire museum in the most prominent position 
in the Milestones of Flight gallery which is at the 
main entrance. What other choice could we have 
made, other than the 1903 Wright brothers flying 
machine, the world's first airplane, catalogued, inci- 
dentally, as item number 100 among those 300 aircraft. 

The story of that machine, and the two men, Wilbur 
and Orville Wright, who designed, built, and flew it, 
has been told many times. The sequence of events 
leading up to the birth of powered flight has justifiably 
received more study than any other episode in the 
history of technology. 

Yet, for all this attention, the fundamental tech- 
nological achievements embodied in the design and 
construction of that machine are still little understood. 
We find it easy to speak of the engineering genius of 
Wilbur and Orville Wright, but have difficulty ex- 
plaining that genius in precise terms. 

On December 16, 1983, as part of the National Air 
and Space Museum's commemoration of the eightieth 

anniversary of the invention of the airplane, we invited 
five distinguished engineers to explore with us specific 
elements of the aerodynamic, structural, and power 
plant technology incorporated in the 1903 Wright 
airplane. As we had hoped, the speakers were able 
to illuminate the approach, methods, and accomplish- 
ments of Wilbur and Orville Wright, drawing on their 
own years of experience in aeronautics. 

We offer these papers to you in this volume, edited, 
annotated, illustrated, and with additional collateral 
material. I think you will agree that the result is far 
more than a retelling of the old story; it throws new 
light upon the epoch-making event of 1903 and the 
genius of the two bicycle manufacturers whose pa- 
tience, persistence, and creativity changed the mo- 
mentum of Mankind's eternal pursuit of greater mo- 
bility throughout its environment. 

Walter J. Boyne is Director of the National Air and Space 
Museum. He was formerly Curator of Aeronautics and Chief 
of Preservation and Restoration at the Museum's Garber 

He was born in East St. Louis, Illinois, in 1929, obtaining 
his first taste of flying at the Parks Air College Field. He 
was commissioned in the U.S. Air Force in 1952 and retired 
as a Colonel in 1974. During his service career he became 
a Command Pilot, with over 5,000 hours of flying time in 
a score of different aircraft. He began writing of his primary 
interest, aviation history, in 1962, and since then has 
published more than 300 articles, as well as eight books. 
He is the associate editor of two national magazines, and 
writes columns for two others. 

Mr. Boyne graduated with honors from both the Univer- 
sity of California, Berkeley (B.S., B.A.), and the University 
of Pittsburgh (M.B.A.). Mr. Boyne serves as National 
Advisor of the American Aviation Historical Society, and is 
on numerous advisory boards for air and space organizations. 


The Wright Brothers 

The First True Aeronautical Engineers 



The year 1983 was the eightieth anniversary of the 
first successful heavier-than-air, powered, manned 
flight. This historic event, accomplished by Wilbur 
and Orville Wright on the cold, windswept sand dunes 
of Kill Devil Hill just outside of Kitty Hawk, North 
Carolina, took place at approximately 10:35 a.m. on 
December 17, 1903. This success was the culmination 
of seven years of hard work, intense dedication, and 
exceptional engineering insight on the part of the 
Wrights. One purpose of this paper is to highlight 
this engineering insight, and to explain why the 
Wrights succeeded where many others failed. After 
1908, when the Wrights first publicly demonstrated 
their flying machine (Wilbur in France and Orville at 
Fort Meyer in Virginia), the advancement of aviation 
took off exponentially. The development of the 
airplane, and subsequently its transmutation into 
space vehicles such as the Apollo lunar vehicle and 
the space shuttle, represents one of the most important 
contributions of the twentieth century, and will con- 
tinue to impact all of world society and mankind for 
the duration of civilization on this planet. There are 
hundreds of thousands of engineers and scientists 
today who can trace the origins of their jobs directly 
back to the Wright brothers, and many millions of 
others — pilots, airline personnel, travelers, military 
planners and practitioners — whose lives are touched 
and frequently dominated by this wonderous inven- 
tion and development of the "flying machine" by the 
Wrights. Among other aspects, the profession of 

aeronautical engineering was born with the success of 
the Wrights. Orville and Wilbur Wright were essen- 
tially the first true aeronautical engineers in history; 
therefore, a second purpose of this paper is to set 
forth the reasons for this distinction. 

Who invented the concept of the modern airplane? 
The answer, amazingly enough, is not the Wright 
brothers. Orville and Wilbur Wright invented the first 
practical airplane, but the concept of the modern 
airplane as a machine with fixed wings, a fuselage, 
and horizonal and vertical tails was introduced long 
before the Wrights were born. Indeed, the Wrights 
inherited a certain technology base in aeronautics that 
had been accumulating for centuries. Another purpose 
of this paper is to survey pre-Wright aeronautics and 
to assess its impact on the ultimate success of the 

Aeronautical Technology before the Wrights 

For a moment, imagine that you have been born and 
raised on an isolated island in the South Pacific, 
completely devoid of all modern knowledge and 
technology. However, through the observation of the 
flight of birds, let us say that you have developed an 
intense interest in flight. Indeed, you want to fly 
yourself. How are you going to do it? Are you going 
to conceive a machine with fixed wings, fuselage, and 
a tail, with some type of separate mechanism for 
propulsion? Not likely! Rather, you would be more 


inclined to directly emulate the birds. Most probably 
you would fashion wings out of feathers, leaves and 
branches, or wood, attach them to your arms, climb 
a hill or a roof, and jump off, flapping wildly. Indeed, 
this is precisely what ancient man did. Early history 
is full of accounts of such attempts. Aviation historians 
have labeled these intrepid people as "tower jump- 
ers." They were uniformly unsuccessful. And, most 
likely, so would be you in your attempts. 

After a number of bumps and bruises, to say the 
least, you might decide to try a different tactic. You 
might design a machine which, by means of pushing 
and pulling levers with your legs and arms, would 
mechanically flap wings for possible lift and propul- 
sion. Such devices were designed in the past and are 
called "ornithopters." Perhaps the most spectacular 
early ornithopter designs were conceived by Leonardo 
da Vinci (1452-1519), an example of which is shown 
in Figure 1. This sketch is taken directly from da 
Vinci's notes, which were only recently discovered in 
the late nineteenth century. Unfortunately, omithop- 
ters have little or no redeeming aerodynamic qualities 
and have been uniformly unsuccessful for manned 

1. An ornithopter design by Leonardo da Vinci, 1486 
to 1490. 

By this time in your flying endeavors, frustration 
would most likely set in, as it did for many early 
would-be aviators. A major glimmer of hope occurred 
on November 21, 1783, when, for the first time in 
history, human beings left the ground in a sustained 
flight through the air. The machine was a hot air 
balloon designed and built by two French brothers, 
Joseph and Etienne Montgolfier. On that date, the 
Montgolfier balloon carried two passengers, Pilatre de 
Rozier and the Marquis d'Arlande for a 25 minute 
drifting flight across the spires and roofs of Paris. 
Although the advent of the hot air balloon, and the 
hydrogen-filled balloon after it, did much to spark the 


2. The first aerial voyage in history: The Montgolfier 
hot air balloon lifts from the ground near Paris, 
November 21, 1783. 

general public's interest in flight (indeed, the balloon 
craze of the late eighteenth century can be seen in 
the enormous number of balloon scenes appearing on 
porcelain, patterns, and paintings from that period), 
it did little to advance the technology for heavier- 
than-air flight. 

How then did the concept of the modern airplane — 
a machine with fixed wings, fuselage, and tail, with 
separate mechanisms for the production of lift and 
propulsion — arrive? The answer, in this author's opin- 
ion, is through a stroke of genius on the part of one 
man — George Cayley (1773-1857). Cayley was a rela- 
tively well-to-do baronet who lived on an estate in 
Yorkshire, England. An educated man, Cayley spent 
his life working intensely on engineering, social, and 
political problems in England. The dominant interest 
of his life was heavier-than-air flight, and in 1799 he 
set forth for the first time in history the concept of 
the modern airplane. He inscribed this concept on a 
silver disk (Figure 4). At the left, on one side of the 
disk, we see a flying machine consisting of a large 
fixed wing, a fuselage below the wing, and a cruciform 
tail. For propulsion, Cayley was enamored with the 
idea of paddles — a misguided thought which is not 

John D. Anderson, Jr. 


5. Cayley's sketch of his whirling arm apparatus, 

3. Sir George Cayley (1773-1857). 

6. Cayley's 1804 glider. 

4. The silver disk on which Cayley engraved his 
concept for a fixed wing aircraft, 1799. 

7. Cayley's "boy carrier," 1849. 

important in light of his overall concept of the sepa- 
ration of the generation of lift from the generation of 
propulsion. On the flip side of this disk (the right of 
Figure 4), we see for the first time in history, an 
aerodynamic force diagram on a fixed lifting surface 
at some inclination angle to the oncoming flow. Here, 
Cayley has clearly identified the drag vector (parallel 
to the flow) and the lift vector (perpendicular to the 
flow). Therefore, the concept of the modern airplane 
was first advanced by George Cayley in 1799. It was 
this concept which was to be parlayed by the Wright 
brothers into the first successful airplane more than a 
century later. 

George Cayley did a lot more for aeronautics than 
just dabble with silver disks. He was the first person 

to mount an intelligent, well-conceived program in 
aeronautical research. In 1804, he built a whirling arm 
apparatus (Figure 5). At the end of this whirling arm 
is a lifting surface (a portion of a wing) on which 
Cayley measured the lift force. Also in 1804, he 
designed, built, and flew a small model glider (Figure 
6). In today's context, such hand-launched gliders are 
child's play, but in 1804 it represented the first modern 
configuration airplane in history, with a fixed wing, 
and a horizontal and vertical tail that could be adjusted. 
During 1809 and 1810, Cayley published three papers 
on his aeronautical research where he quite correctly 
pointed out for the first time that: (1) lift is generated 
by a region of low pressure on the upper surface of 
the wing; and (2) cambered wings (curved surfaces) 

The Wright Brothers 

generate lift more efficiently than a flat surface. These 
results, among many others, can be found in his 
papers entitled "On Aerial Navigation" published in 
the November 1809, February 1810, and March 1810 
issues of Nicholson's Journal of Natural Philosophy. 
This "triple paper" by Cayley ranks as one of the 
most important aeronautical documents in history. 
Cayley went on to put theory into practice. In 1849, 
he designed, built, and tested a full-size airplane. A 
triplane (Figure 7), which during some of its tests 
carried a ten-year-old boy through the air several 
meters above the ground while gliding downhill. For 
this reason, the machine is sometimes called "Cayley's 
boy carrier." One of Cayley's other designs (Figure 
8) appeared in Mechanics' Magazine in 1852. Cayley 
never achieved his final goal — sustained heavier-than- 
air, powered, manned flight. However, he is clearly 
the grandfather of the modern airplane. (For an entire 
book on Cayley and his aeronautical contributions, 
see Reference 1.) 

Cayley's concept for an airplane was fixed in the 
public's eye by a design carried out by William Samuel 
Henson. Henson's aerial steam carriage embodied 
Cayley's concepts of a fixed wing, fuselage, and tail. 
Moreover, it was powered by a steam engine driving 
two rotating propellers — a propulsion concept infi- 

-Ail ROW, SEPTEMBER 2.'., 1852. 'Snut ii„ St.noptd id. 


Pig. 2. 

9. Henson's aerial steam carriage, 1842-1843. 

10. Stringfellow's triplane, 1868. 

nitely more viable than Cayley's flappers. Henson 
was a contemporary of Cayley's, and his ideas for the 
steam carriage were published in 1842. The machine 
was never built, but prints such as that shown in 
Figure 9 were widely distributed around the world, 
and served to crystallize in the mind of the general 
public what an airplane — whenever it would be suc- 
cessful — might look like. 

Also contemporary with Cayley, and a friend of 
Henson's, was John Stringfellow, who made several 
attempts to bring Henson's design to fruition. For 
example, he built a model triplane (Figure 10), which 
was exhibited in London in 1868. Stringfellow de- 
signed several steam engines for aeronautical use, but 
all were simply too heavy for the power they produced. 

Adding to the list of attempts and failures during 
the last quarter of the nineteenth century were those 
of Felix Du Temple and Alexander F. Mozhaiski. 
Du Temple was a Frenchman who designed and built 
a flying machine with forward-swept wings (Figure 
11). In 1874, at Brest, France, this aircraft achieved 

8. Cayley's 1852 glider. 

11. Du Temple's airplane, 1874. 

John D. Anderson, Jr. 

12. Mozhaiski's airplane, 1884. 

the world's first "powered hop" when it was launched 
down an inclined plane and left the ground for a 
moment. Ten years later, Mozhaiski in Russia de- 
signed and built the aircraft shown in Figure 12. It 
too was launched down an inclined plane and flew 
for a few seconds. It represented the second "powered 
hop" in history. However, neither flight represented 
anything close to the sustained, controlled flight 
necessary for success. 

The last half of the nineteenth century witnessed 
a growing list of serious technical achievements in 
aeronautics. For example, in 1870 the Frenchman 
Alphonse Penaud was the first to experiment with 
rubber-band-powered model aircraft, with which he 
studied stability and control aspects associated with 
wing dihedral and various tail incidence angles. One 
of his model aircraft, the "planophore," is shown in 
Figure 13. A promising aeronautical career was cut 



14. Francis H. Wenham (1824-1908). 

gested by Richard Harte in England in 1870 (Figure 
15). Harte conceived this device to counteract the 
torque of a propeller and to act as a differential drag 
mechanism to steer an airplane in directional flight. 

13. Penaud and his planophore, 1870. 

15. Harte's ailerons, 1870. 

short by Penaud's suicide in 1880. Also during this 
period, the Englishman Francis H. Wenham discov- 
ered that most of the lift on an airplane wing was 
obtained from the portion near the leading edge. As 
a result, in 1866, Wenham was the first person to 
discuss the benefits of high aspect ratio wings. Wen- 
ham went on to design, build, and use the first wind 
tunnel in history, in 1871. In another development, 
the use of a flaplike surface at the trailing edge of a 
wing — today we would call it an aileron — was sug- 

However, he did not think of this aileron as a roll 
control device for inducing motion about the airplane's 
longitudinal axis. The stunning aspect of roll control 
was left to the Wright brothers to invent. Also during 
this period the first serious airfoil development was 
carried out by the Englishman Horatio F. Phillips. 
Using the second wind tunnel in history, Phillips 
conducted numerous experiments on double surface 
airfoils (Figure 16). These airfoil shapes were patented 
by Phillips — the six upper shapes in 1884 and the 

The Wright Brothers 

16. Airfoils of Horatio F. Phillips. 

lower airfoil in 1891. The Aeronautical Society of 
Great Britain, formed in 1866, helped to add cohesion 
and credibility to this aeronautical research. The 
membership constituted serious researchers with good 
scientific reputations. It sponsored regular scientific 
meetings and technical journals. The Society still 
flourishes today in the form of the highly respected 
Royal Aeronautical Society, and has been the role 

model for subsequent societies such as the modern 
American Institute of Aeronautics and Astronautics in 
America today. 

An interesting figure just before the turn of the 
century was Sir Hiram Maxim, an American expatriate 
from Texas living in England. Maxim designed and 
built a huge steam-powered machine (Figure 17), 
which perhaps best illustrates the culmination of 
aeronautical technology of the late nineteenth century. 
Although it never successfully flew, on July 31, 1894, 
it did rise above a guide track and ran for a length of 
600 feet before being snagged by a guard rail several 
feet above the track. Maxim's aircraft is a perfect 
example of what the famed aviation historian Charles 
H. Gibbs-Smith termed the "chauffeur" philosophy. 
Simply stated, this philosophy embodied all efforts 
to get into the air by brute force: give me an engine 
powerful enough to lift the machine into the air, and 
I will not worry too much about how to fly it once 
airborne. I will simply drive it around like a carriage 
on the ground. This philosophy was uniformly un- 

In contrast, the "airman's" philosophy was quite 
different. Here, the approach is to first get up in the 
air, fly around with gliders, and obtain the "feel" of 
an airplane before putting an engine on the vehicle. 

17. Sir Hiram Maxim and his airplane, 1894. 


John D. Anderson, Jr. 

This "airman's" philosophy was first advanced by the 
German engineer Otto Lilienthal, who designed and 
flew the first successful gliders in history. Lilienthal 
collected a large bulk of aerodynamic data, which he 
published in 1889 in a book entitled Der Yogelflug ah 
Grundlage der Fliegekunst (Bird flight as the basis of 
aviation). This book greatly influenced aeronautical 
design for the next 15 years, and was the bible for 
the early designs of the Wright brothers. During the 
period 1891 — 1896, Lilienthal made over 2,500 suc- 
cessful glider flights. With these, he advanced the 
cause of aeronautics by leaps and bounds. He was a 
man of aeronautical stature comparable to Cayley and 
the Wright brothers. Unfortunately, during a glider 
flight on August 9, 1896, Lilienthal stalled and crashed 
to the ground. His spine was broken, and he died a 
day later in a Berlin hospital. There is some feeling 
that had Lilienthal lived, he might have beaten the 
Wright brothers to the punch. No one will ever know; 
however, it is certain that the Wright brothers picked 
up the "airman's" philosophy from Lilienthal, and as 
we will soon see, this philosophy ultimately led to 

As a final note in this section, we mention again 
that Lilienthal was killed in 1896. The year 1896 is 
a "red-letter" date in the history of aviation for several 

other reasons, which will soon be apparent. Please 
keep 1896 in mind as you read further. 

The Race for the First Flight — Langley and the 
Wright Brothers 

By the last decade of the nineteenth century, aero- 
nautical technology had matured far beyond the early 
"tower jumpers" and ornithopter designs, and indeed 
by this time the concept originally set forth by George 
Cayley in 1799 had been refined to the extent that 
success had to be imminent. This technology base 
developed during the 1800s was directly inherited by 
two groups working at the turn of the century in the 
United States. One group was led by Samuel Pierpont 
Langley at the Smithsonian Institution, and the other 
constituted two young brothers in Dayton, Ohio — 
Wilbur and Orville Wright. Whether they realized it 
or not, these two groups were in a race for the first 
powered flight. This section tells their story. 

Examine Figure 19. What you see is a photograph 
of a confident, self-assured, almost pompous man. He 
is Samuel P. Langley, who in the year 1896 repre- 
sented, after Lilienthal, probably the most serious 
and potent aeronautical researcher and designer in 
the world. Langley was born in Roxbury, Massachu- 

18. Otto Lilienthal and one of his gliders. 

The Wright Brothers 

19. Samuel P. Langley (1834-1906). 

setts, in 1834, and graduated from Boston High School 
in 1851. He consciously decided not to go to college; 
rather, he joined an architectural firm in Boston so as 
to learn civil engineering and architecture in a practical 
way. He later educated himself in astronomy, and 
began to travel extensively in Europe. After working 
a year as an assistant in the Harvard College Observ- 
atory in Cambridge, Langley was made an assistant 
professor of mathematics at the Naval Academy in 
Annapolis. One year later, in 1867, he became director 
of the Allegheny Observatory in Pittsburgh. Langley 
was obviously an upwardly mobile man, and while at 
the Allegheny Observatory, he made a reputation for 
himself as a leading expert on sun spots. 

In 1885, Langley attended a meeting of the Amer- 
ican Association for the Advancement of Science in 
Buffalo, New York. There he heard a paper on the 
flight of soaring birds and the technical possibility of 
manned flight. This paper had a major impact on 
Langley, and it kindled an interest in flight that 
burned within Langley's soul to the day he died. 
Immediately after this meeting, Langley requested 
and obtained permission from the trustees of the 
Observatory to build an aeronautical laboratory. He 
quickly constructed a whirling arm apparatus with 
which he began to measure the aerodynamic properties 
of lifting surfaces, much like George Cayley 80 years 
earlier. This work initiated a decade of intense aer- 
onautical research by Langley. Most of this work, 


> V 

20. Langley's whirling arm apparatus, 1887. 

however, was carried out at the Smithsonian Institu- 
tion in Washington, D.C., because in 1887 Langley 
accepted an offer to become the Secretary of the 
Smithsonian — the most prestigious scientific position 
in nineteenth-century America. At the Smithsonian, 
in the original brick building now labeled "the Cas- 
tle," Langley built an aerodynamic laboratory. There 
he experimented with nearly a hundred different 
model airplane configurations, a few of which are 
shown in Figure 21. This work culminated in a steam- 
powered, unmanned, heavier-than-air machine which 
he called an "aerodrome." It was a tandem-wing 
design, about 16 feet in length. Langley mounted 
this aerodrome on a catapult on top of a houseboat 
and, on May 6, 1896, on a wide portion of the Potomac 
river near Quantico, Virginia, the aerodrome was 
launched. It flew under its own power for one and a 
half minutes, covering a distance of over a half mile. 
It was the first successful sustained flight of a heavier- 
than-air, powered vehicle in history. A smashing 
accomplishment for which Langley deserves more 
accolades than modern aviation history usually accords 
him. Again, on November 28 of the same year, the 
aerodrome flew almost two minutes, covering approx- 
imately a mile. Elated with his success, Langley felt 
he had accomplished his goal — to demonstrate the 
feasibility of powered flight. After ten years of hard 
effort, extensive research, and considerable expend- 
iture of resources and money, Langley wrote the 

John D. Anderson, Jr. 


-^J^r C5»«o 





21. Langley's airplane models. 

22. Langely's small-scale aerodrome, 1896. 


I have brought to a close the portion of the work 
which seemed specially mine. For the next state, the 
commercial and practical development of the idea, it 
is probable that the world may look to others. 

Note the year of Langley's success, 1896 — the same 
year as Otto Lilienthal's death. 

Also in this same year, 1896, Wilbur Wright became 
serious about powered flight. He shortly interested 
his brother Orville in an endeavor which was to become 

23. Langley's aerodrome mounted on the houseboat, 1896. 

The Wright Brothers 



■ Ay |g 

24. Langley's small-scale aerodrome in flight. These 
pictures were taken by Alexander Graham Bell, 
inventor of the telephone and close friend of 
Langley's, 1896. 

the consuming passion for the rest of their lives. 
Wilbur (born in 1867) and his younger brother, Orville 
(born in 1871), were members of a family which put 
much emphasis on intellectual achievement, a dedi- 
cated work ethic, high ethical behavior, and modesty. 
These personal traits, present in Wilbur and Orville 
to an almost excessive amount, are among the primary 
reasons for the ultimate success of the Wrights. Their 
mother had a college degree in mathematics, a very 
unusual achievement for that day. Their father was a 
bishop in the United Brethren Church, a very accom- 
plished man who had published books and articles, 
and who held various administrative positions in the 

church. As a matter of interest, neither Wilbur nor 
Orville officially received a high school diploma: 
Wilbur was not able to attend his commencement 
exercises (a requirement to obtain his degree), and 
Orville took a special series of courses in his senior 
year that did not lead to a prescribed degree. Like 
Langley, they did not have a formal college education; 
they were self-made men. 

The Wright brothers began their aeronautical work 
by reading all the available literature, thereby absorb- 
ing the aeronautical technology described in the 
previous section. In particular, they concentrated on 
a book entitled Progress in Flying Machines, published 
in 1894 by Octave Chanute, a widely respected 
American civil engineer who became an intense pro- 
ponent of powered flight during the last part of the 
nineteenth century. The Wrights also became familiar 
with the details of Otto Lilienthal's aerodynamic data, 
and used these data in their first two glider designs. 
In 1899, after observing the soaring flight of large 
birds, Wilbur conceived the idea of bending, or 
deflecting, the tips of wings to achieve lateral control 
around the longitudinal axis of an aircraft. This 
concept, popularly called wing warping, was another 
one of the major ingredients for the Wrights' success. 

The Wrights were airmen in the great tradition of 
Lilienthal. In 1900 they designed, built, and experi- 
mented with a glider at Kitty Hawk, North Carolina. 
This aircraft (Figure 27) had a 17-foot wingspan, a 

25. Wilbur Wright (1867-1912). 

26. Orville Wright (1871-1948). 


John D. Anderson, Jr. 

27. Wright brothers' 1900 glider. 

28. Wright brothers' 1901 glider. 

29. Manly (left) and Langley. 

personal backing) offered Langley a $50,000 contract 
to build a single man-carrying, powered flying ma- 
chine. Langley accepted. He hired as an assistant 
Charles Manly, a graduate fresh from Cornell's Sibley 
School of Mechanical Engineering. For all practical 
purposes, Langley theorized that he needed only to 

horizontal elevator in front of the wings, and was 
usually flown on strings from the ground. Only a few 
brief piloted flights were made, mainly because the 
glider did not perform to the Wrights' expectations. 
The glider contained the best state-of-the-art aero- 
nautical technology for that day, but it simply did not 
fly right. It produced far less lift than calculated by 
the Wrights, and, therefore, had to be flown at a 
higher angle-of-attack and/or in the face of higher 
winds than expected. In 1901, the Wrights returned 
to Kitty Hawk. They had redesigned their glider, this 
time making it larger with a 22-foot wingspan. Again, 
the new glider embodied the best aeronautical state- 
of-the-art, and again the Wrights were not satisfied 
with its performance. Indeed, on the train returning 
to Dayton after their 1901 trials, Wilbur is quoted as 
stating in a moment of despair that "nobody will fly 
for a thousand years." 

Meanwhile, back in Washington, Langley's self- 
imposed retirement from aeronautical research was 
short-lived. In 1898, spurred by the Spanish-American 
War, the War Department (with President McKinley's 

30. Manly's engine, the first radial engine in 
aeronautical history. 

The Wright Brothers 


scale his previously successful aerodrome to full size. 
However, he wisely made the decision to switch from 
steam power to a gasoline-fueled internal combustion 
engine. Moreover, he calculated that, as a minimum, 
he needed an engine that could produce 1 2 horsepower 
and weigh 100 pounds or less — a horsepower to weight 
ratio larger than any existing engine of that day. 
Langley gave a subcontract to the Balzer Company 
in New York to design and manufacture such an 
engine — the company went bankrupt trying. Manly 
then personally took over the engine design, and by 
1901 had assembled a radically designed five cylinder 
radial engine that weighed only 124 pounds and 
produced a phenomenal 52.4 horsepower. This engine 
(Figure 30) was to be the best airplane power plant 
designed until the middle of World War I. In June 
1901, Langley tested a quarter-scale gasoline-pow- 
ered, unmanned aerodrome, which flew perfectly 
(Figure 31). At this stage, Langley was literally and 
figuratively "cooking with gas." 

■ , 

31. Langley's gasoline-powered quarter-scale 
aerodrome, 1901. 

Meanwhile, back at Dayton, the Wrights overcame 
their despair about their disappointing 1901 glider 
performance. At this stage, they made a decision 
which took great intellectual courage. To this date, 
they had faithfully relied upon detailed aerodynamic 
data published by both Lilienthal and Langley. Now 
they wondered about its accuracy. Wilbur wrote that 
"having set out with absolute faith in the existing 
scientific data, we were driven to doubt one thing 
after another, until finally, after two years of experi- 
ment, we cast it all aside, and decided to rely entirely 
upon our own investigations." This momentous change 
in philosophy is one of the major reasons for the 
Wright brothers' ultimate success. Besides courage, 
it took a great deal of self-confidence — hallmarks of 
the Wrights. Between September 1901 and August 
1902, the Wrights undertook a major program of 
aeronautical research. They built a wind tunnel in 
their bicycle shop in Dayton, where they tested more 
than 200 different airfoil and wing shapes, some of 
which are shown in Figure 33. This period of aero- 
nautical research was a high-water mark in early 
aviation development. With their new-found data, 

32. Wright brothers' wind tunnel, 1901-1902. 

the Wrights designed a new glider, and in 1902, at 
Kill Devil Hill, this machine flew very successfully. 
Figure 34 is one of many photographs taken of their 
successful 1902 machine — it shows a graceful flight 
of the most successful glider developed to that point 
in history. Note the degree of wing warping clearly 
evident. During 1902, the Wrights made over 1,000 
glider flights. They set a distance record of 622.5 feet 

No. 311 31 34 

33. Wright brothers' airfoil and wing models, 1901- 


John D. Anderson, Jr. 

34. Wright brothers' 1902 glider. 

and a duration record of 26 seconds. Moreover, they 
had control about all three axes of the aircraft, which 
allowed them to make smooth, banked turns. In the 
process, both Wilbur and Orville became highly skilled 
and proficient pilots, each achieving more than an 
hour and a half of flight time. The Wright brothers 
were carrying the "airman" philosophy to its natural 
conclusion. Flushed with success, the Wrights re- 
turned to Dayton in the fall of 1902, ready for the 
final step — the inclusion of a power plant. 

Once again, the Wrights did everything themselves. 
Splitting responsibilities, Wilbur tackled the problem 
of the propeller, and Orville set about to obtain an 
engine. At first thought, Wilbur assumed that he could 
find the necessary design information on propellers 
in the Dayton library — after all, steamships had been 
using water screws for almost a century. He was 
disappointed to find only empirical design data which 
applied to ships, and which were useless for aeronau- 
tical application. This situation prompted Wilbur to 
do his own work. In the process, he developed the 
first rational airplane propeller theory in history. He 
was the first to recognize that a propeller is nothing 
more than a twisted wing, where the "lift" force is 
now pointing forward for propulsion. Using his theory 
in conjunction with their airfoil data measured during 
the previous year, Wilbur designed and constructed 
a remarkably efficient propeller. This aspect of the 
Wright brothers' technology is sometimes not fully 
appreciated, yet it was one of the most important 
technical victories that led to their success. 

In parallel, Orville's first effort on the engine was 
to inquire with the existing motor manufacturers about 
the possibility of delivering an engine that could 

produce 12 horsepower and weigh less than 150 
pounds. The responses were uniformly negative. 
Hence, following in their tradition, Orville, with the 
aid of their bicycle mechanic, Charlie Taylor, designed 
and built an original engine that just met their 
minimum specifications. It was not nearly as impres- 
sive as Charles Manly's engine for Langley's aerod- 
rome, but it was sufficient for the job. 

With their engine and propeller development be- 
hind them, the Wright brothers constructed a new 
machine from scratch during the summer of 1903. 
Popularly called the Wright Flyer, it closely resembled 
the 1902 glider, but had a larger wingspan of 40 feet, 
4 inches, and used a double rudder behind the wings 
and a double elevator in front of the wings. A 
photograph of this machine, with its engine and twin 
pusher propellers, is shown in Figure 35. This pho- 

35. The 1903 flying machine, popularly called the 
Wright Flyer. 

The Wright Brothers 


tograph was taken during the early fall of 1903 at Kill 
Devil Hill, where the Wrights were making prepara- 
tions for a powered flight. 

Meanwhile, back at the Smithsonian, Langley was 
also making preparations. His full-scale aerodome was 
readv. Manly had volunteered to be the pilot, over 
Langley's objections. Langley's original plans were 
to first test the aerodrome with a lifeless dummy 
aboard, but Manly successfully argued that too much 
money and time had already been spent on their 
project, and that the schedule for a manned flight 
could be delayed no longer. Therefore, on October 
7, 1903, Langley's aerodrome, mounted on a bigger 
and better houseboat in the Potomac, was ready for 
flight. The launching was given wide advance pub- 
licity, and the press was present to watch what might 
be the first successful flight in history. What happened 
next is described in the resulting report as it appeared 
in the Washington Post the next day: 

A few yards from the houseboat were the boats of 
the reporters, who for three months had been 
stationed at Widewater. The newspapermen waved 
their hands. Manly looked down and smiled. 
[Author's note: Why Manly was smiling is not easy to 
understand. Here was a man, ready to fly, with zero 
flight time. Both Langley and Manly were 
"chauffeurs." But let us go on with this newspaper 
account.] Then his face hardened as he braced 
himself for the flight, which might have in store for 
him fame or death. The propeller wheels, a foot 
from his head, whined around him one thousand 
times to the minute. A man forward fired two 
skyrockets. There came answering "toot, toot" from 

36. Langley's full-scale aerodrome on the houseboat, 


the tugs. A mechanic stooped, cut the cable holding 
the catapult; there was a roaring, grinding noise — and 
the Langley airship tumbled over the edge of the 
houseboat and disappeared in the river, sixteen feet 
below. It simply slid into the water like a handful of 
mortar .... 

37. Langley's first launch, October 7, 1903. 

Manly was unhurt. Langley believed the airplane 
was fouled by the launching mechanism, and he tried 
again on December 8, 1903. Again, Manly was at the 

38. Langley's second launch, December 8, 1903. 

John D. Anderson, Jr. 

controls. A dramatic photograph taken a moment after 
this second launch shows the aerodrome going through 
a 90 degree angle-of-attack, with its rear wings totally 
collapsed. Again, Manly was fished out, fortunately 
unhurt. But this was the end of Langley's attempts. 
In light of today's massive government-sponsored 
aerospace research and development program, it is 
important to note that the $50,000 War Department 
contract given to Langley at the turn of the century 
was the first of its type. Moreover, Langley had 
utilized an additional $23,000 in Smithsonian re- 
sources to help design and construct the aerodrome. 
As a result, Langley's failures had a rather adverse 
impact on the political thinking of that day towards 
government-sponsored research. For example, James 
A. Hemenway, chairman of the House Appropriations 
Committee, stated: 

If it is to cost us $73,000 to construct a mud duck 
that will not fly 50 feet, how much is it going to cost 
to construct a real flying machine? 

Representative Robinson of Indiana had this to say: 

Langley is a professor, wandering in his dreams of 
flight, who was given to building castles in the air 
.... If steps are not taken to curb research 
spending, someone will influence some Department 
to test the principle of erecting buildings beginning 
with the roof. 

Representative Hitchcock of Nebraska, when inter- 
viewed by a Brooklyn newspaper, exclaimed: 

You tell Langley for me that the only thing he ever 
made to fly was Government money! 

Finally, from the official War Department report on 
the entire project: 

We are still far from the ultimate goal, and it would 
seem as if years of constant work and study by 
experts, together with the expenditure of thousands 
of dollars, would still be necessary before we can 
hope to produce an apparatus of practical utility on 
these lines. 

Nine days after Langley's second failute, the Wright 
Flyer rose from the sands of Kill Devil Hill. 

In retrospect, the vilification of Langley by the 
press, the public, and government officials was much 
too harsh. Langley died in 1906, a broken man. He 
had made a valiant try, and he had exerted all his 
scientific knowledge and insight to accomplish his 
goal. He simply had not been able to design the right 
system — something that the Wright brothers under- 
stood much more fully than Langley. However, the 
successful flights of Langley's small aerodrome rep- 
resented major aeronautical accomplishments for that 
period. For these, Langley should be given due credit. 

Let us turn again to the Wright brothers. After the 
development of their propellers and engine, and after 
constructing their new machine, the Wrights returned 
to Kill Devil Hill in September 1903. Due to a 

combination of bad weather and engine shaft failure, 
they were delayed far into the fall. Wilbur and Orville 
had originally planned to fly their new machine first 
as an unpowered glider in order to get the feel of its 
flying qualities. However, with their long delays, and 
due to their intense desire to fly that year, they made 
a decision somewhat uncharacteristic of their normally 
conservative, careful approach: they decided that their 
first flight would be powered. 

Their first opportunity finally arrived on December 
14. The brothers flipped a coin to see who should be 
the first pilot. Wilbur won. The Flyer, under power, 
took off, suddenly went into a steep climb, stalled, 
and thumped back to the ground. It was the first 
recorded case of pilot error in powered flight: Wilbur 
admitted that he put on too much elevator and brought 
the nose too high. (Had the Wrights been able to 
follow their original plans to fly the machine as a 
glider first, then most likely Wilbur would not have 
overcontrolled, and the first successful powered flight 
would have gone on the record books on December 
14 with Wilbur as the pilot.) A photograph of the 
aircraft just after Wilbur's attempt shows some slight 
damage to the canard surfaces in front of the wing 
(Figure 39). 


39. Wilbur Wright, moments after his first attempted 
powered flight, December 14, 1903. 

This damage was repaired in a matter of days, and 
the Flyer was again ready for flight on December 17. 
This time it was Orville's turn at the controls. It was 
a cold, windy day, almost too windy for flying. Again, 
the Wrights were taking an uncharacteristic risk. At 
10:35 a.m. the restraining rope was released, and the 
machine began to move. Figure 40 tells the rest of 
the story. It is the most famous photograph in the 
history of aviation. It shows the Wright Flyer at the 
moment of liftoff, with Wilbur running along side to 
keep the wing tip from dragging in the sand. The 
epoch-making flight lasted for 12 seconds and covered 
120 feet. It was, in Orville's words, "the first in the 
history of the world in which a machine carrying a 

The Wright Brothers 


40. December 17, 1903. 

man had raised itself by its own power into the air in 
full flight, and sailed forward without reduction of 
speed, and had finally landed at a point as high as 
that from which it started." There were three more 
flights that day, the last remaining in the air for 59 
seconds and covering 852 feet. The world of powered, 
manned, heavier-than-air flight — and along with it the 
world of successful aeronautical engineering — had 
been born. 

Concluding Remarks 

Wilbur and Orville Wright were indeed the first true 
aeronautical engineers in history for the following 

1. They were the first to fully recognize the impor- 
tance of flight control around all three axis of the 
airplane. Their concept of wing warping for lateral 
control was pioneering — it distinguished them from 
all prior inventors, and set them ahead of all contem- 
porary aviators for a decade. 

2. They were the first to modify and improve their 
flight controls by means of a systematic series of 
successful glider flights in 1902. In this vein, they 
might even be labeled the first true test pilots as well. 

3. They were the first to use wind tunnel results to 

correct some defective data existing in the literature. 
Moreover, although they never tested a full airplane 
configuration in their tunnel, their wind tunnel models 
were directly aimed at optimizing a specific flight 
configuration. Therefore, the Wrights were the first 
to use a wind tunnel in the modern mode that we 
see today. 

4. The Wrights' understanding of the true aerody- 
namic function of a propeller, and their subsequent 
development of a propeller theory are important firsts 
in the history of aeronautical engineering. This work 
sets them apart from any previous investigators. 

5. The fact that the Wrights, with no prior experience 
with any type of internal combustion engine, were 
able to design and build a successful engine that was 
beyond the state-of-the-art, and to accomplish this in 
the space of six months, is truly amazing. 

6. Finally, and probably most importantly, the Wrights 
were the first to treat a flying machine as an integrated 
system involving aerodynamics, propulsion, structures, 
and flight dynamics. They fully appreciated the in- 
teraction and mutual importance of all these aspects. 
In this sense, they were the first to build a total flying 
machine — a machine which had all the major aspects 
that a modern airplane has today. 


John D. Anderson, Jr. 

The above reasons, based on their technology, 
account for the Wright brothers' success. However, 
there are other, more philosophical reasons why the 
Wright brothers succeeded where others failed. First, 
they had an intense work ethic which was brought 
into focus day and night, day after day, towards their 
goal. Second, they were supremely self-confident, 
never really doubting the value of their final goal. 
Third, they were intellectually courageous, willing to 
do things on their own when the existing technology 
either failed them, or did not measure up to their 
requirements. Moreover, they had the insight to 
recognize good technology from bad technology. Fi- 
nally, without the benefit of a formal education, the 
Wright brothers had an innate sense of engineering — 
very systematic minds that logically and inexorably 
led them to success. 

This article has set the stage for the papers that 
follow. In these papers, you will have the opportunity 
to probe deeply into the technology of the Wright 
brothers. You will be amazed to observe how sound 
that technology was. You will be swamped with the 
realization that their technology was not by chance or 
happenstance — they were truly the first aeronautical 

As a final note, if you are interested in reading 
more extensively about the early development of 
aeronautics from antiquity to the Wright brothers, 
there are numerous books that treat the subject. 
References 1-8 are just a few of the many which you 
may find of particular interest. It is from these books 
that much of the material in this article was gleaned. 


1. Gibbs-Smith, Charles H. Sir George Cayley's Aeronautics: 
1796-1855. London: Her Majesty's Stationery Office, 

2. Aviation: An Historical Survey from its Origins to the End 
of World War II. London: Her Majesty's Stationers- 
Office, 1970. 

3. Josephy, A.M., and Gordon, A. The American Heritage 
History of Flight. New York: Simon and Schuster, 1962. 

4. McFarland, Marvin W., ed. The Papers of Wilbur and 
Orville Wright. New York: McGraw-Hill, 1953. 

5. Anderson, John D., Jr. Introduction to Flight: Its 
Engineering and History. New York: McGraw-Hill, 1978. 

6. Crouch, Tom D. A Dream of Wings. New York: W. W. 
Norton & Co., 1981. 

7. Hallion, Richard P., ed. The Wright Brothers: Heirs of 
Prometheus. Washington, D.C.: Smithsonian Institution 
Press, 1978. 

8. Combs, Ham', with Cardin, Martin. Kill Devil Hill. 
Boston: Houghton Mifflin Co., 1979. 

9. Vaeth, J. Gordon. Langley: A Man of Science and Flight. 
New York: Ronald Press, 1966. 

John D. Anderson, Jr. is a leading engineering educator. 
A native of Lancaster, Pennsylvania, he is a graduate of the 
University of Florida and holds a Ph.D. in Aeronautical and 
Astronautical Engineering from Ohio State University. He 
has served as a task scientist with the U.S. Air Force at the 
Aerospace Research Laboratory at Wright Patterson AFB 
and as Chief of the Hypersonics Group, Aerophysics Di- 
vision at the Naval Ordnance Laboratory at White Oaks, 
Maryland. Since 1973 he has been a Professor of Engineering 
at the University of Maryland and was Chairman of that 
Department from 1973 to 1980. He has published over 70 
technical papers and is the author of three books. 

Dr. Anderson is a member of Tau Beta Pi, Sigma Xi, 
Sigma Tau, American Physical Society, and the American 
Society for Engineering Education. He is an Associate 
Fellow of the American Institute of Aeronautics and As- 
tronautics and a Fellow of the Washington Academy of 
Sciences. In 1975, he was chosen an Outstanding Educator 
of America, and won the Engineering Sciences Award of 
the Washington Academy of Science. He is listed in Who's 
Who in America and has been designated by the University 
of Maryland as a Distinguished Scholar/Teacher. 

The Wright Brothers 


Aerodynamics, Stability, and 
Control of the 1903 Wright Flyer 



The design, construction, and flight of the 1903 Wright 
Flyer was a scientific engineering achievement of the 
first order. ' It is true, as the Wright brothers thoroughly 
appreciated, that their first powered flights were really 
only an intermediate success. They worked for two 
more years to improve their design until they had a 
practical airplane. But it is proper that we celebrate 
December 17, 1903, as the beginning of aviation. By 
then the Wrights had in hand practically all of the 
fundamental understanding and knowledge they needed 
to show the world how to fly. 

Even by modern standards, the Wright brothers' 
program was extraordinarily well-conceived and effi- 
ciently executed. They conducted the necessary tests, 
collected only the data they needed, and generally 
conducted their work to learn just what they required 
to succeed. Other papers in this collection will treat 
the Wrights' work on engines and structures. We 
restrict our discussion here to aerodynamics, stability, 
and flight control. 

The following pages amount to a progress report 
covering contributions by many people. In 1953 mem- 
bers of the Los Angeles Section of the American 
Institute of Aeronautics and Astronautics (AIAA) con- 
structed a reproduction of the 1903 Flyer. That 
airplane was destroyed in the fire at the San Diego 
Aerospace Museum in 1977; shortly after that event, 
the Los Angeles AIAA section received the insurance 
claim. Mr. Howard Marx of the Northrop Corporation, 
as chairman of the AIAA committee on special events, 
proposed that a flying reproduction be constructed. 

The idea was enthusiastically supported and the AIAA 
Wright Flyer Project was born. We set out more than 
five years ago with dozens of people to do by com- 
mittee what the Wright brothers alone did in less than 
four years! And we still haven't flown our "Flyer"! 

Our plans have expanded. We now intend to 
construct two reproductions. One is an accurate full- 
scale rendition of the 1903 Flyer to be tested in a 
wind tunnel. It is complete except for covering (Figure 
1). The flying reproduction will incorporate small 
changes from the original design to make the aircraft 
easier to fly safely. Much of the material covered in 
this paper will serve as the basis for determining those 
changes. Equally important is our effort to interpret 
the Wrights' accomplishments in terms of the knowl- 
edge we have gained in the 80 years since their first 

We shall describe some of the results obtained from 
wind tunnel tests of two models, a Ve scale model 
tested at the California Institute of Technology, and 
a Vs scale model tested in a high speed tunnel whose 
owners will identify themselves at some later date. 
The data have been analyzed, partly with the help of 
some theoretical calculations at the Douglas Aircraft 
Company, to provide firm assessments of the stability 
and control of the 1903 Flyer. Using modern control 
theory, analyses carried out at Systems Technology, 
Inc. have helped us understand how the aircraft 
actually behaved when the Wright brothers flew it. 
The results are particularly interesting for the contro- 
versial interconnected wing warp/rudder devised by 


1. Uncovered full-scale 1903 Flyer. (Los Angeles AIAA Wright Flyer Project, 1983) 

the Wrights for lateral directional control. 

It was not a good airplane but it was by far good 

The Wrights' Wind Tunnel Data 

Probably the best-known scientific work by the Wrights 
is their program to obtain data for airfoils and wings. 
Theirs was not the first wind tunnel — which was 
invented in England, by Wenham and Browning in 
1871 (Reference 3) — nor were theirs the first wind 
tunnel data obtained in the United States. Albert C. 
Wells measured the correct value for the drag coeffi- 
cient of a flat plate, reported in his thesis submitted 
to the Massachusetts Institute of Technology in 1896 
(Reference 4). Wells converted a ventilation duct for 
his work; the Wrights designed and built a small open 
circuit tunnel. With that device, during three months 
in 1901 they took the first extensive systematic data 
suitable for the design of aircraft. The results served 
them well for a decade. 

Ten years earlier, Otto Lilienthal had used a 
whirling arm apparatus to measure the lift and drag 
for various airfoils approximating the shape of birds' 
wings (Reference 5). The Wright brothers used his 
data in the design of their 1900 and 1901 gliders. It 
is a familiar fact that because they obtained substan- 

tially less lift with their gliders than they had predicted 
with Lilienthal's results, the Wrights resolved to 
obtain their own data. What is less well known is that 
in the course of their program they determined that 
Lilienthal's data were essentially correct. 

The difficulty lay with the value of a coefficient 
which was required to convert Lilienthal's numbers 
to obtain the actual aerodynamic forces acting on a 
wing. That coefficient — the drag force acting on a 
unit area of plate oriented perpendicular to a stream 
moving with speed one mile per hour — was called 
Smeaton's coefficient. 

John Smeaton was the preeminent English civil 
engineer of the eighteenth century. In 1752 he pub- 
lished an important memoir (Reference 6) in which 
he discussed theory and experiment for the fluid 
mechanics of water wheels and windmills. He in- 
cluded a table of data, provided by a Mr. Rouse, from 
which the coefficient defined above can be deduced 
and shown to be approximately 0.0049, independent 
of velocity. Thus the drag on a plate having area S in 
a stream moving at speed v (MPH) is 

D = 0.0049V 2 S 

The value 0.0049 is for air, being proportional to the 
density of the medium. Presumably because of Smea- 
ton's stature and because he authored the book, his 


F.E.C. Culick and Henry R. J ex 

name was subsequently attached to this number. Mr. 
Rouse, who actually did the work, has hardly ever 
since been cited. 

In any case, this value of Smeaton's coefficient 
persisted for 150 years. The strength of tradition 
caused Lilienthal to accept the value without question. 
But the Wrights determined otherwise. With a clever 
combination of their wind tunnel data and a few tests 
with a wing from their 1901 gliders, they concluded 
that the correct value was 0.0033, which is now known 
to be correct for the range of speeds in which they 
were working. Samuel P. Langley (Reference 7) had 
previously found this result, confirmed later by Wells. 

Figure 2 shows the close agreement between the 
measurements of Lilienthal and those of the Wrights 
for the same parabolic airfoil. They are expressed 
here in the modern terms, lift coefficient (lift force 
divided by the dynamic pressure and area) as a function 
of the angle of incidence between the flow and the 
airfoil. The shift of the Wrights' view from their initial 
belief that Lilienthal's data were seriously in error, to 
the recognition that their own results agreed with his, 
is a superb illustration of the objective and thoroughly 
professional fashion in which they carried out their 
work. The following selections from Wilbur's diary 
(Reference 1) summarize the development of their 

October 6, 1901 

I am now absolutely certain that Lilienthal's table 
is very seriously in error, but that the error is not so 
great as I had previously estimated . . . . If in our 
Kitty Hawk calculations we had used a coefficient of 
.0033 instead of .005 the apparent advantage of our 
surface over the plane as per the Duchemin formula 
would have been much greater. I see no good reason 
for using a greater coefficient than .0033. 


^-Lilienthal (1889) 


\ \ 
\ \ 


/ ^^ \ 

Lift -8 


/ ^^^~-~*. 



t- Wright (1901) 

C L .6 











i i i i 

-10 10 20 30 40 

Angle of Incidence, a 


2. Data for lift coefficient versus angle of incidence, 
Lilienthal and the Wright brothers. 

October 16, 1901 

It would appear that Lilienthal is very much nearer 
the truth then we have heretofore been disposed to 

November 2, 1901 

Lilienthal is a little obscure at times but, once 
understood, there is reason in nearly all he writes. 

December 1 , 1902 

The Lilienthal table has risen very much in my 
estimation since we began our present series of 
experiments for determining lift . . . for a surface as 
near as possible like that described in his book the 
table is probably as near correct as it is possible to 
make it with the methods he used. 

Thus the Wrights concluded that Lilienthal's data 
were correct and that the cause of their low prediction 
of the lift force was the incorrectly high value of 
Smeaton's coefficient. They never measured the cor- 
rect value directly, but deduced it from their wind 
tunnel tests for an airfoil and their small number of 
measurements for a full-scale wing. Their reasoning, 
experimental work, and results are all truly remarka- 
ble. They are especially impressive when one realizes 
that this effort was motivated entirely by the practical 
need to obtain information necessary to the successful 
design of their aircraft. This is a very early example 
of a process which is now so common that it is taken 
for granted. The demands of an engineering program 
may pose a question which can be satisfactorily 
answered only by fundamental scientific work com- 
pleted outside the main thrust of the engineering 
effort. It was one of the great strengths of the Wrights 
that they were able to identify, formulate, and solve 
crucial basic problems. In contrast, their contempor- 
aries trying to build flying machines were able to 
progress only with crude trial-and-error methods of 
traditional nineteenth-century engineering and inven- 
tion. With their philosophy and style the Wright 
brothers were solidly in the twentieth century, far 
ahead of their contemporaries in aviation. That is the 
major reason for their rapid and certain progress to 
manned flight. 

Fundamental Notions of Stability 

Nothing related to the Wright brothers has created 
more confusion, controversy, discussion, and at times 
vitriolic argument than questions of equilibrium, sta- 
bility, and control. There is fairly general agreement 
that the Wrights' experience with bicycles taught 
them the virtue of control. The bicycle is unstable 
without active control by the rider. Thus the Wrights 
were not deterred by the possibility of an unstable 
vehicle which could nevertheless be successfully op- 
erated with practice, providing the means existed for 
proper control. It is also clear that control was always 
a central issue during development of their aircraft. 2 

Aerodynamics, Stability, and Control 


What is by no means evident is the extent to which 
the Wrights inadvertently produced unstable aircraft. 
They certainly refused to follow their contemporaries 
who were preoccupied with the goal of inventing an 
intrinsically or automatically stable airplane. On the 
other hand, it is not necessary that an airplane be 
unstable to be controllable. 

The Wrights were the first to place the smaller 
horizontal surface forward — the canard configuration. 
They knew very well the history of the aft horizontal 
tail. In particular, they were aware that, as perceived 
by Cayley in 1799 and shown by Penaud in 1872, an 
aircraft with an aft tail can be made longitudinally 
stable. Moreover, early in their program, in 1899 with 
the kite, and in 1900 with the man-carrying kite/ 
glider, they experimented successfully with an aft 
tail. They knew that the configuration could easily 
be made stable. There is no doubt that they chose 
the canard because of fear, first expressed by Wilbur, 
that the aft tail carried with it an intrinsic danger. 
What worried them was the possible inability to 
recover from a stall, due to loss of lift induced by a 
vertical gust, or by the pilot upon raising the nose too 
far. That had been the cause of Lilienthal's death in 

At least twice during the tests in 1901, Wilbur 
found himself in a stalled condition. By manipulating 
the canard he was able to get the nose down and the 
aircraft mushed to the ground without serious damage. 
He was therefore convinced that his reasoning was 
correct. A certain sense of security was given the pilot 
because he was able to see the actions of the control 
surface, which also provided a visual reference relative 
to the ground. 

Thus the choice of the canard configuration, the 
most distinctive feature of the Wright aircraft, was 
not based on sound technical grounds of stability. It 
was rather a matter of control in pitch, especially 
under extreme conditions. In fact, the Wrights did 
not understand stability in the precise sense that we 
do now. The reason is fundamental: nowhere in their 
work did they consider explicitly the balance of 
moments. 3 They shared that ignorance with all others 
trying to build aircraft at that time. So strictly, whether 
their aircraft were stable or unstable was an accidental 
matter. Often, changes in a design were made which 
would change the stability, and not always favorably. 
But the motivation was always the desire to affect 
some observable characteristic, such as undulations 
in pitch. From this point of view, the question of the 
Wrights' intentions to design an unstable airplane is 

For our later discussion of the wind tunnel data, it 
will be helpful to understand the ideas of equilibrium 
and stability. When an aircraft is in steady motion, 
there must be no net force or moment acting. For 

horizontal flight, the vertical lift must exactly equal 
the weight and the thrust of the propulsion system 
must be just sufficient to overcome the drag. The 
symmetry of the aircraft guarantees that there shall 
be no net side force. 

In order that there be no net moment tending to 
rotate the aircraft, the moments about three axes must 
separately vanish: the pitch, roll, and yaw moments 
must all vanish for equilibrium. Much work is saved 
in practice by using coefficients rather than the mo- 
ments themselves. A moment coefficient is obtained 
by dividing the moment by the dynamic pressure; 
the wing area; and a length, either the wing chord 
for the pitching moment or the wing span for the roll 
and yaw moments. The moment coefficients are given 
the symbols C|, C m , and C n for roll, pitch, and yaw, 
respectively, as shown in Figure 3. 

To ensure equilibrium or trim, the moment coef- 
ficients must vanish, C, = C m = C n = 0, static 
condition. Whether or not the equilibrium state is 
stable depends on the changes of the aerodynamic 
moments when small disturbances are applied to the 
aircraft. Consider an aircraft in steady horizontal flight. 
Suppose that a vertical gust causes an increase in the 
angle of incidence between the flow and the aircraft. 
The initial equilibrium state may be restored if the 
increased incidence generates a pitching moment 
causing the nose to pitch down so as to reduce the 
angle of incidence to its initial value. By convention, 


Pitching Moment 
Yaw Moment 
Roll Moment 


C m =0 
Cn = 


When a disturbance is appliedfeg. a gust) the change 
of aerodynamic moment must be such as to restore 


^Sj, Roll Axis 

Yaw Axis 

3. Basic conditions for equilibrium and stability. 


F.E.C. Culirk and Henry R. J ex 

X Pitching Moment 

a) Pitch 

Cm =-08 

Yaw Moment 


Angle of Sideslip 

b) Yaw 

c) Roll 
4. The three basic moment curves. 

a pitching moment tending to rotate the nose down 
is defined to be negative. The preceding reasoning 
shows that for stability of equilibrium, the pitching 
moment must decrease when the lift increases. This 
behavior is sketched in the upper portion of Figure 
4. The lift is plotted versus the pitching moment, 
with negative pitching moments to the right of the 
vertical axis. 4 For stable equilibrium the pitching 
moment curve, shown dashed in the sketch, must 
slope from the lower left to upper right and intersect 
the lift axis; at that point, the pitching moment is 
zero and small displacements along the curve are 
accompanied by changes of the pitching moment 
tending to restore the equilibrium state. 

The solid curve labeled unstable also passes through 
the equilibrium point, but small displacements cause 
changes of pitching moment which act to increase the 
displacement. The curve has been drawn to pass 
through a point labeled — .08, which we shall see later 
is the value of the pitching moment for zero lift of 
the 1903 Flyer. The original Flyer was very unstable 
in pitch. Note that a stable pitching moment curve 
can obviously be drawn through an equilibrium point 

(zero moment) requiring negative lift. 

We can apply similar reasoning to motions in yaw, 
with the result sketched in the middle of Figure 4. 
If the nose of the aircraft is disturbed to the left of 
the path, the wind strikes the right side and the 
aircraft is slipping to the right; this is by definition a 
positive angle of sideslip. For directional stability, a 
positive (nose to the right) yaw moment must be 
generated, causing the nose to swing to the right into 
the wind. Hence the curve of yaw moment versus 
angle of sideslip must slope up to the right for stability. 
Directional stability is provided mainly by the vertical 
tail; the 1903 Flyer had acceptable, though not large, 
directional stability. 

Finally we consider stability in roll, commonly 
called "dihedral" effect. The main idea is that if a 
wing drops, a rolling moment will eventually be 
generated to restore the wing's level. If, for example, 
the right wing drops, gravity causes the aircraft to fall 
to the right, producing a positive angle of sideslip. 
This motion must then create a negative rolling 
moment lifting the right wing. If the dihedral effect 
is positive, the curve of roll moment versus angle of 
sideslip must therefore slope downward to the right 
as sketched in the lower portion of Figure 4. Positive 
or upward dihedral angle of the wings produces 
positive dihedral effect. Thus, the opposite condition, 
negative dihedral effect, is sometimes called "anhed- 
ral" effect. This was used by the Wrights in their 
1903 Flyer. 

To summarize, flight in stable equilibrium requires 
that six conditions be satisfied. For equilibrium, the 
three moments about the pitch, roll, and yaw axes 
must vanish. For the equilibrium to be stable, changes 
of the moments produced by small deviations from 
the equilibrium state must act to restore the initial 
state. Application of this requirement has shown what 
slopes the moment curves must have for a stable 

In this general context we have treated equally the 
rotational motions about the three axes. Motions in 
pitch hold a special position, however, owing to 
fundamental characteristics of the usual aircraft having 
a longitudinal plane of symmetry. In steady level 
flight, the plane of symmetry is vertical and contains 
both the direction of flight and of gravity. The pitch 
axis is perpendicular to the plane of symmetry and 
rotations in pitch directly affect the vertical motion. 
A fundamental and general property of the pitch 
stability of aircraft must be emphasized. It is always 
true that moving the center of gravity forward will 
make an airplane more stable, for the following reason. 
When an airplane is in flight, application of an 
aerodynamic moment, whether by action of the con- 
trols or due to an atmospheric disturbance, causes 
rotation about an axis passing through the center of 

Aerodynamics, Stability, and Control 


gravity. Consider the case of a vertical gust, which 
causes the angle of incidence to increase, so the lift 
is increased. Imagine that the center of gravity is very 
far forward, ahead of all lifting surfaces. Then clearly 
an increase of the lift forces on the wing and tail 
produce a rotation forcing the nose down, tending to 
decrease the angle of incidence. This is a stable 
response. Similarly, if the center of gravity is aft of 
all lifting surfaces, an increase of angle of incidence 
will be further encouraged, the change of lift forces 
causing the airplane to pitch up. This is an unstable 
reaction. It is reasonable to expect that somewhere 
between the unrealistically extreme locations there 
should be a position of the center of gravity for which 
the aerodynamic forces generate no net pitching 
moment in response to a distrubance of the angle of 
incidence. That location of the center of gravity is 
called the neutral point (NP) — every airplane has one. 
For a conventional airplane, the neutral point is 
somewhere on the wing chord, perhaps 30 percent- 
40 percent aft of the leading edge. For a canard 
configuration the neutral point is much closer to the 
leading edge, and often lies ahead of the wing. 
Positions of the neutral points are labeled NP in 
Figure 5(a-b). 

Note •' Arrow length denotes local Q_ (lift /area) 

a) Stable (e.g. ahead of neutral point) 

• Center of gravity forward 

• Forward surface stalls first —pitch down 

• Recovery '."automatic"; control with aft 
surface (unstalled) 

I) Aft Toil (Penaud) 

2) Canard (Rut an) 

b) Unstable (e.g. behind neutral point ) 

• Center of gravity aft 

• Aft surface stalls first -—pitch up 

• Recovery: control with forward surface (unstalled) 

3) Aft Tail -Relaxed 
Stability (Birds) 

4) Canard (Wrights) 

5. Stable and unstable wing/tail configurations. 


Longitudinal Stability of Aft Tail and Canard 

In his classic paper describing his rubber-band-pow- 
ered model airplane (Reference 8), Penaud gave the 
first detailed analysis of longitudinal or pitch stability. 
It was not a general discussion; the main purpose had 
been to show how an aft tail can stabilize pitch 
motions. The limited scope seems subsequently to 
have helped create some misunderstanding. For ex- 
ample, it has often not been appreciated that just as 
a configuration with aft tail is not necessarily stable, 
so also a canard configuration (which Penaud did not 
consider) may be stable or unstable. A correct theory 
of the stability of all cases did not appear until 1904 
in the seminal paper by Bryan and Williams (Ref- 
erence 9). 

A wing alone can be made stable, but only if 
particular care is taken to use a proper airfoil shape 
having a reflexed camber line. This seems to have 
been realized first by Turnbull in 1906 (Reference 
10). However, a flying wing brings its own problems 
and we need consider here only the more common 
case of a main wing and a smaller horizontal surface 
for stabilization and control. Four cases are possible: 
the smaller surface is either forward of aft of the wing, 
and each of those configurations may be stable or 

The four are shown in Figure 5, with labels citing 
the best-known examples of each. The lengths of the 
arrows in Figure 5 represent the relative loads per 
unit area or lift coefficient; C b when the configuration 
is trimmed for equilibrium in pitch. This shows the 
most important distinction between stable and un- 
stable configurations. Whatever the relative sizes of 
the surfaces, the forward surface carries more load per 
unit area when the configuration is stable: the value 
of its lift coefficient is greater than that for the aft 
surface. As a result, if the angle of incidence is 
increased, the forward surface will usually stall first. 
This means that for a conventional stable aircraft 
(Figure 5.1), the wing stalls first and may lose lift 
suddenly, but the aft tail continues to be effective 
and can be used to control pitch motions. In particular, 
the tail can be used to generate a nose-down moment, 
causing the wing to recover its lift. When the lifting 
forward surface of a stable canard stalls (Figure 5.2), 
the nose drops, but while the canard is stalled, precise 
pitch control is not possible. 

An unstable aircraft having an aft tail (Figure 5.3) 
can be extremely difficult, if not fatally dangerous for 
man to fly, although soaring birds often fly in this 
condition. The most critical condition again arises 
with the behavior at high angles of incidence. Now 
the aft tail may stall before the wing, control is lost, 
and the wing stalls soon after. The possibility of 

F.E.C. Culick and Henry R. J ex 

operating such configurations successfully, and thereby 
gaining their advantage of increased efficiency, can 
be realized with the use of automatic controls. This 
is a subject of growing interest and application in 
modern aircraft design. 

And so we arrive at the final case, Figure 5.4, the 
unstable canard used by the Wright brothers (and 
rarely since!). If the angle of attack is sufficiently 
high, the aft surface, now the main lifting surface, 
may stall first. While this appears to be extremely 
serious, the saving grace is that, unlike the previous 
case, control is not lost. And that is probably why the 
Wrights were successful with their unstable gliders' — 
they always had control. If the wing has large camber, 
as with the Wrights' 1903 airfoil, the canard must 
carry additional lift to balance the large diving pitching 
moment due to the wing. As a result, the canard may 
stall first as the angle of attack of the aircraft is 
increased. That seems to have been the case for the 
1903 Flyer as we shall show later. 

For our wind tunnel data we estimate that the 
neutral point of the 1903 Wright Flyer was about 10 
percent of chord aft of the leading edge. The center 
of gravity was 30 percent aft of the leading edge, so 
the airplane was severely unstable. The difference of 
those two numbers, —20 percent or —.20 is called 
the static margin. For current aircraft with automatic 
control, the greatest negative static margin which is 
acceptable is about —5 percent. 

It follows from the discussion of stability and the 
neutral point that the slope of the curve lift coefficient 
versus moment coefficient (or simply lift versus pitch- 
ing moment) depends on the location of the moment 
reference point, the position of the center of gravity. 

If the center of gravity is moved aft from a stable 
location, the slope tends to be less upward to the 
right, becoming more upward to the left. The curve 
must pass through the value of the residual pitching 
moment at zero lift, so the moment curves become 
skewed as shown in Figure 6. Here we have used the 
data taken with the Vb scale model discussed in the 
following section. The position of the center of gravity 
for which the curve is vertical is the neutral point; for 
these data, the neutral point is at approximately 0.10 
times the wing chord (c), or 10 percent of the chord. 

Vortex Lattice Calculation of Aerodynamics 

As a part of the AIAA Wright Flyer Project, two 
members of the aerodynamics committee have used 
modern computational techniques to calculate some 
of the major aerodynamic characteristics of the aricraft. 
Using two different vortex lattice computer programs, 
James Howford and Stephen Dwyer of the Douglas 
Aircraft Company have calculated load distributions, 
lift and pitching moment for the Flyer. We believe 
that these are the first such analyses of the aircraft 
and in fact may be the first applications of vortex 
lattice theory to a biplane! 

The main idea of vortex lattice theory' is that the 
influences of an object in a flow can be calculated by 
replacing that object by a distribution of vorticity over 
its surface. Vorticity is an elementary form of fluid 
motion which can be visualized as a collection of 
microscopic vortices or whorls — little tornadoes side- 
by-side. Figure 7 shows how the airplane is treated 
for this purpose. The wings, canard, and vertical tail 
are approximated as surfaces having zero thickness, 

At C L 

Wrights' eg 


CL \\\\\Y\\\\v 


Total of 300 cells 

6. Influence of the center of gravity on the pitching 
moment curve. 

7. Approximation to the Wright Flyer for vortex 
lattice calculations. 

Aerodynamics, Stability, and Control 


not a bad assumption for the 1903 Flyer. For these 
calculations the surfaces have been divided into 300 
panels, over each of which the vorticity is constant. 
The procedure requires solving 300 equations for the 
300 values of vorticity or loading on the panels. No 
account is taken of the struts, truss wires, and other 
structure external to the load-carrying surfaces. In the 
vortex lattice method the flow is assumed to be inviscid 
so the drag is zero. The drag due to lift, the induced 
drag, can be calculated but is not included here. 

Examples of Howford's load distributions are given 
in Figure 8. The loading per foot of span on the lower 
wing is plotted for several conditions. Figure 8(a-c) 
shows the influence of canard deflection. In part (a) 
the load distribution has the nearly elliptical form 
expected for changes of incidence for the wing alone. 
Deflection of the canard (nose up) produces downwash 
behind the canard and upwash in the region outside 
its tips. This produces a negative loading in the central 
portion of the wing, and a slight increase in the 
outboard regions, part (b). The net loading on the 
wing for changes of both canard and wing incidence 
is shown in part (c). 

In part (d) and (e) of Figure 8 the incremental 
loadings on the wing due to pitch and yaw rates are 
illustrated. The wake of the canard has a large 
influence in pitch, and relatively less in roll. 

Not shown here, but evident in the results of the 


a) Wing -a lone 

J I I L 

b) Canard 


cJ Angle -of -attack 
(canard and wing) 

d) Pitch Rate 

e) Roll Rate 


8. Load distributions calculated with vortex lattice 

vortex lattice calculations, is the significant upstream 
influence of the wing. The spanwise loading on the 
wing produces a strong upwash field decaying within 
several wing chord lengths. Because the canard is 
located within the upwash field, this aggravates the 
contribution of the canard to pitch instability by an 
additional 25 to 30 percent. 

These results show directly the obvious fact that 
the flow induced by the canard may have substantial 
effects on the lift generated by the wing, and vice 
versa. This feature cannot be ignored in analysis of 
the aerodynamics of the Flyer. Suitable integration 
of results such as these will give the total lift and 
moment for the aircraft. The accuracy of the calcu- 
lations will become apparent upon comparison with 
data taken in wind tunnel tests. 

Results and Interpretation of Wind Tunnel 

We have carried out two series of wind tunnel tests 
within the AIAA Wright Flyer Project. The first used 
a Vfc scale model shown in Figure 9. The tests were 
carried out in the GALCIT ten-foot tunnel at the 
California Institute of Technology (Reference 11). 
Because one of the main intentions of the tests was 
to obtain data for the effectiveness of wing warping, 
the model was built of wood and fabric, with steel 
truss wires, very similar to the original aircraft. As a 
result, the structure was relatively fragile and suffered 
considerable damage during the test program. Some 
of the results seem to be biased because of distortions 
of the wing surfaces. 

The second series of tests used the stainless steel 
model, Vs scale, shown in Figure 10 (Reference 12). 
Extensive tests were carried out, including changes 
of configuration to investigate possible modifications 
for the full-scale flying reproduction mentioned ear- 
lier. An advantage of the steel model is that data can 
be taken at higher speeds, or Reynolds numbers. The 
Reynolds number for the tests ranged from 50 to 90 
percent of the value in full-scale flight. In this range 
the aerodynamic properties suffer only small changes. 

Figure 1 1 is a sketch of the profile of the aircraft 
showing the definition of several quantities that are 
important in presenting the data. We have chosen the 
reference location of the center of gravity to be 30 
percent aft of the leading edge of the lower wing and 
30 percent of chord above the lower wing. This choice 
is based on estimates by Professor Hooven of Dart- 
mouth College and by Mr. Charles McPhail of the 
AIAA Wright Flyer Project. The bottom of the skid 
rail is the horizontal reference. A line drawn through 
the centers of the leading edge and the aft spar is 
parallel to the skid line; this defines the angle of zero 
incidence of the upstream flow. The same reference 


F.E.C. Culick and Henry R. J ex 

9. Covered V& scale model in the GALCIT ten-foot tunnel, California Institute of Technology. 

10. Stainless steel Vh scale model 

Aerodynamics, Stability, and Control 



(chord line through centers of LE and TE) 

Wright 03 Flyer 4.07° 

Steel WT model 4.60° 

Covered wood model 3.5 


11. Profile of the 1903 Flyer showing reference lines and center of gravity. 


Vortex -lattice theory ( Dwyer 1981) 

O — o Steel model at RN = .90 x I0 6 ( Heglund 1983) 
xxx Covered model at RN=.43x I0 6 (Cutick 1982) 






_C L =.62 //////// y//&C D =.\\7 /////////////////////// '!jl * f m ///////////$ 




i i i i i 

-1 1 L3| I I 

04 .16 .20 .24 .28 -12 

Drag Coefficient, Cn 

12. Lift and drag of the 1903 Flyer; comparison of theory and tests. 


8 12 


F.E.C. Culick and Henry R. J ex 

line defines the zero angle of canard deflection. 

Lift and Drag Aerodynamics 

Here we shall discuss only a portion of the data, to 
illustrate some comparisons between experiment and 
theory, and to cover some of the results used later in 
calculations of the stability, control, and dynamics of 
the airplane. Figure 12 shows two of the basic char- 
acteristics of an airplane, the drag polar, lift coefficient 
versus drag coefficient; and the lift curve, lift coeffi- 
cient versus angle of attack. Because the steel model 
has larger structural members for strength at the higher 
test speeds, the drag is larger than that for the Vb 
scale model (called covered model) at the lower lift 
coefficients. The horizontal crosshatched line is drawn 
at the value of lift coefficient we estimate to be that 
for cruising flight of the original Flyer. The agreement 
of data for the drag of the two models at this value 
of lift coefficient must be regarded as fortuitous: data 
for drag are often suspect, and especially for these 
models the results may be sensitive to the value of 
the Reynolds number. 

The lift curve slope obtained with the steel model 
is very closely matched by the calculations based on 
vortex lattice theory, showing an angle of incidence 
of about one degree at cruise. This suggests again the 
understanding of aerodynamics possessed by the Wright 
brothers: it appears that the geometrical setting of the 
wing, with respect to the skid rail, was very closely 
that required for cruise flight. The lift curve for the 
covered model has closely the same slope as the other 
two results, but is displaced by roughly four degrees 
to higher angles of attack. This seems to be due to 
an average reduction of the camber of the airfoil due 
to distortion of the structure. In any case, both sets 
of data show that the cruise lift coefficient is well 
below the value for stall of the aircraft, further 
evidence of careful design by the Wrights. 

Pitching Moment Aerodynamics 

A summary of our present understanding of the 
pitching moment of the 1903 Flyer is given in Figure 
13. The best data, those taken with the steel model, 
are displayed as open symbols; results are shown for 
three canard settings, degrees and H — 10 degrees. 
It appears that a deflection of about + 6 degrees (nose 
up) is required for a trim condition having zero pitching 
moment at the cruise lift coefficient of 0.62. But 
according to our earlier discussion of Figure 4, this is 
an unstable condition because the slope of the curve 
lift coefficient versus moment coefficient is downward 
to the right. 

The data taken with the Vb scale covered model 
are plotted as the crosses. These show a smaller value 
of pitch down pitching moment at zero lift. Corre- 
spondingly, the elevator deflection for trim is nose 

down, producing a pitch down moment on the air- 
plane. The smaller pitching moment at zero lift is 
consistent with the smaller angle of incidence for zero 
lift shown by the data in Figure 12. Both deficiencies 
may be explained by somewhat less camber or a small 
amount of symmetrical twist (trailing edge up) of the 
wings on the covered model. It appears that the 
second may be the more likely explanation — unless 
the data for the steel model and the result of the 
vortex theory are both. in error! 

Whatever the case, it is best not to try to "correct" 
the data, a practice universally understood now, but 
less well recognized in the Wrights' time. In a letter 
to Chanute, Wilbur offered the following astute ob- 
servation concerning Langley's treatment of some of 
his own data for lift on a flat plate: "If he had followed 
his observations, his line would probably have been 
nearer the truth. I have myself sometimes found it 
difficult to let the lines run where they will, instead 
of running them where I think they ought to go. My 
conclusion is that it is safest to follow the observations 
exactly, and let others do their own correcting if they 
wish." (Reference 1, p. 171). We follow Wilbur's 
dictum and present both sets of our wind tunnel data. 

The unstable pitching characteristic of the 1903 
Flyer is arguably its worst feature, although as we 
shall see, the lateral characteristics are also poor. The 
large negative static margin ( —20 percent) meant that 
the airplane was barely controllable. Three compen- 
sating factors made the flights on December 17 
possible: the low speed, high damping of the pitching 
motions, and most importantly the Wrights' living 
skills. During their development work leading to the 
1905 airplane, the first practical airplane, the brothers 


Vortex- lattice 


/ (Dwye 



— o 


model at 

RN = 


s (Hegl 

and 1983) 

© © 



ed model 

at RN =.43 x 

l0 6 (Cu 

ick 1982) 

Elevotor Settings- 
S.= io° 0° 

.12 .08 .04 
Pitching Moment 
Coefficient, C mcG 

Static Margin : 

Ax nr . rr , = -.20C (unstable) 

13. Pitching moment of the 1903 Flyer; comparison 
of theory and tests. 

Aerodynamics, Stability, and Control 


made two important changes: they increased the area 
of the canard, and they added weight, as much as 
70-140 pounds, along the upper longeron to the 
canard, to bring the center of gravity forward (Ref- 
erence 13). 

Those improvements were made to ease the diffi- 
culties they encountered controlling undulations in 
pitch, a dynamical consequence of the static instability 
we have been examining. In fact, the most significant 
cause of the unstable pitch characteristic is the large 
negative pitching moment at zero lift (see Figure 12). 
Referring to Figure 4, we see that in order to be able 
to trim an aircraft for a condition of stable equilibrium, 
it is necessary that the pitching moment at zero lift 
be positive and at trim lift be zero. The Wrights 
achieved the latter by moving the center of gravity 
aft, but they could not then satisfy the former. 

The large negative pitching moment at zero lift of 
the 1903 Flyer is due almost entirely to the airfoil. A 
highly cambered airfoil must operate at a relatively 
large negative angle of incidence to produce zero lift. 
At that condition the pressure distribution is such that 
a large negative (nose down) pitching moment is 
generated. This is easily demonstrated qualitatively — 
hold a curved plate in an airstream. It is possible that 
the Wrights were aware of this behavior, but it is 
more likely that they were not. Nowhere do they 
discuss the pitching moment characteristics of airfoils. 
We have already remarked that they were apparently 
unaware of the necessity for using the dynamic equa- 
tions for moments to obtain a thorough and correct 
understanding of stability. 

So the Wrights followed Lilienthal and used thin, 
highly cambererd airfoils resembling the cross-sections 
of birds' wings. They were misled to believe that 
airfoils of that sort produced the highest ratio of lift/ 
drag. There is in fact much truth in this conclusion 
if data are taken from small wings at the low speeds 
the Wrights used in their wind tunnel tests. Thicker 
airfoils having less camber are superior for full-scale 
aircraft. However, it is the large negative pitching 
moment of the Wrights' airfoil that is the main issue. 
Simply by reducing the camber, they could have 
achieved enormous improvement in the longitudinal 
flying characteristics of their aircraft. In their later 
aircraft they apparently reduced the camber, but not 
as much as they could have. 

Directional Aerodynamics 

The data for lateral and directional characteristics of 
the two models, plotted in Figures 14 and 15, seem 
to agree acceptably well. Note that the scales on the 
axes are different (the tests were done and the original 
reports were prepared in two different laboratories). 
The sideforce generated in sideslip, Figures 14a and 
15a, is relatively small because there is practically no 


Coefficient, "j-OI 


3 - 12 16 

Sideslip Angle, /3(deg) 

1 . 

i i i 


C £,s \ 


Sideslip Angle, /3(deg) 
i i . , , i i . . i , , , i 



-8 -4 

4 8^ 12 16 


14. Data for lateral and directional characteristics of 
the 1903 Flyer at trim (Vh scale model). 

side area other than the vertical tail. The slope of the 
curve C n versus (3 is small but positive as it should 
be for directional stability (Figures 14b and 15b). 
According to the shift of the curves — i.e., the change 
of yaw moment with rudder deflection, 8 R — the rudder 
had plenty of trim control effectiveness. A rudder 
deflection of ten degrees gives zero yaw moment for 
a trim angle of sideslip equal to eight degrees. That 
means that in steady flight, 0.8 degrees of sideslip 
can be maintained for each degree of rudder deflec- 
tion. Compare this with a pure vertical tail alone for 
which one degree of rotation would trim at exactly 
one degree of sideslip. 

Lateral Aerodynamics Anhedral 

One of the distinctive features of the 1903 Flyer is 
that the wings are rigged for anhedral — the tips are 
"arched" as the Wrights called it, about eleven inches 
below the centerline. This produces a positive vari- 
ation of roll moment with sideslip which, according 
to our remarks in connection with Figure 4c, is an 
unstable response. Suppose that in steady level flight 
the right wing tip drops. Gravity causes the airplane 

F.E.C. Culick and Henn R- J ex 

Trim /3= + 8 for S v = + IO 

1 1 

. i i 


C i,s 

i , i i 






)ue to 1 

-8f "4 .j 

*l -oi 


^ 8 12 16 
Sideslip Angle, /3(deg) 

right wing 

15. Data for lateral and directional characteristics of 
the 1903 Flyer at trim (Vfc scale model). 

to slip to the right, giving a positive angle of sideslip. 
It is evident that with anhedral, the crosswind tends 
to strike the upper surface of the lowered wing, 
forcing it to fall further. This is an unstable response. 

Thus we see in both Figures 14c and 15c that the 
slope of the data for roll moment versus sideslip is 
positive as expected. The slope is less for the data 
taken with the Vb scale covered model, a result which 
may be at least partly explained by symmetric twist 
which would tend to reduce the anhedral of the aft 
portions of the wing. Both curves are biased so that 
there is a non-zero (negative) value of roll moment 
even with no sideslip. This is due to the fact that the 
right wing has slightly larger span than the left, 
approximately four inches for the full-scale Flyer. 
The Wrights built in this small asymmetry to com- 
pensate the weight of the engine, which was heavier 
than the pilot located on the other side of center. 

The use of dihedral was invented by Cayley some- 
time after 1800. Its purpose is to provide stability in 
roll as described earlier. From the beginning of their 
work, the Wrights chose not to use dihedral. Writing 
to Chanute in February 1902, Wilbur refers to a letter 

by a third party: 

He seems surprised that our machine had a safe 
degree of lateral equilibrium without using the 
dihedral angle. He has not noticed that gliding 
experimenters are unanimous in discarding that 
method of obtaining lateral stability in natural wind 
experiments. (Reference 1, p. 217). 

While others, like Lilienthal, were shifting their 
weight to maintain lateral equilibrium, the Wrights 
were using wing warping, which gave them a great 
deal more control. 

In 1900 and 1901 the Wrights' gliders had anhedral, 
to discourage the natural tendency for the aircraft to 
maintain equilibrium, and to allow more effective use 
of the warp control. Their first glider in 1902 was 
rigged so the wings were straight (Reference 1, p. 
322). But early in their 1902 flying season, the Wrights 
again installed anhedral. The reason was a problem 
they encountered because they were gliding close to 
the surface of sloping ground. Orville wrote in his 
diary in September 1902: 

After altering the truss wires so as to give an arch 
to the surfaces, making the ends four inches lower 
than the center, and the angle at the tips greater 
than that at the center, we took the machine out, 
ready for experiment .... We found that the 
trouble experienced heretofore with a crosswind 
turning up the wing it first struck had been overcome 
and the trials would seem to indicate that with an 
arch to the surfaces laterally, the opposite effect was 
attained. (Reference 1, p. 258). 

What they disliked was the obvious consequence 
of dihedral: if the airplane is exposed, say to a 
crosswind from the right (which is the same as positive 
sideslip), the roll moment which is generated by 
positive dihedral lifts the right wing, as the wind 
"catches" the under surface. When the aircraft has 
low directional stability — as was the case for their 
glider — there is only a weak tendency for the nose to 
turn into the wind. The net effect for their early 
gliders was that the left wing tip was driven towards 
the ground. In an attempt to counteract this motion, 
Wilbur had operated the canard to raise the nose and 
the glider stalled, ending in a crash landing. That is 
the "trouble experienced" mentioned in the above 
quotation, and the reason why the Wrights favored 
anhedral which produces the opposite effect: in re- 
sponse to a gust the airplane automatically rolls away 
from the hill. 

That was fine for short, nearly straight flights in 
gliders at the Kill Devil Hill. The powered flights in 
1903 were too brief to show otherwise. But the Wrights 
discovered during their flight tests of 1904 and 1905 
that anhedral has serious undesirable consequences, 
particularly in turning flight. 

Suppose the right wing drops, so gravity causes the 
aircraft to skip to the right. If the wing has anhedral, 

Aerodynamics, Stability, and Control 


this positive sideslip generates a rolling moment 
tending to lower the right wing further (the crosswind 
produces increased pressure on the upper surface of 
the right wing). That is obviously an unstable se- 
quence of events. If, as usually is the case, the aircraft 
has positive directional stability, the nose will be 
swung into the wind, here to the right. The net result 
is that in a right turn, the right wing continues to 
drop; the aircraft changes heading to the right and 
what begins as a small disturbance develops into an 
unstable spiral. 

The motion just described is an unstable form of a 
fundamental motion called the spiral mode. It is part 
of aircraft dynamic stability, a subject more compli- 
cated than the matters of static stability we have 
discussed so far. For example, an aircraft may be 
stable in roll (positive dihedral effect), but if the 
directional stability is sufficiently large, the spiral 
mode will be unstable. Thus, although the aircraft is 
statically stable in the sense shown in Figure 4, it is 
dynamically unstable. That is, in fact, commonly true 
of full-scale aircraft. 

We shall discuss the dynamics of the 1903 Flyer in 
the following section using modern techniques of 
analysis. The Wrights learned the hard way, by flight 
tests, that anhedral aggravated the spiral instability 
with dangerous consequences when they tried to turn 
the aircraft. Although we are here concerned mainly 
with the 1903 Flyer, it is interesting to learn what 
the Wrights did about anhedral in their later aircraft. 
In September 1904 they began practicing turns, at- 
tempting a full circle first on September 15. They 
succeeded on September 20. Then on September 26, 
Wilbur noted in his diary that Orville had been "unable 
to stop turning." The same entry appears on October 

15, but this time the aircraft suffered serious damage. 
"Unable to stop turning and broke engine and skids 
and both screws, Chanute present." On the same 
day, Chanute noted in a memorandum, "Wright 
thinks machine arched too much as speed too great 
across the wind." Thus they seem to have correctly 
located the problem as the anhedral causing the spiral 
mode to be so unstable as to make controlled turning 
extremely difficult. 

After removing the anhedral, the Wrights began 
flying on October 26. The first flight again ended 
with damage to the aircraft. Referring to this incident 
in a letter to Chanute on November 15, Wilbur noted 
"that changes made to remedy the trouble which 
caused Orville's misfortune gave the machine an 
unfamiliar feeling, and before I had gone far I ran it 
into the ground and damaged it again. On November 
2nd we circled the field again, and repeated it on the 
3rd. On the 9th we went out to celebrate Roosevelt's 
election by a long flight and went around four times 
in 5 minutes 4 seconds." Photographs of the airplane 
with anhedral (August 13) and without anhedral 
(November 10) are reproduced here as Figures 16 and 
17. 5 

Although they were able to turn, success was 
intermittent. In fact, the day after he wrote to Chan- 
ute, Wilbur remarks in his diary, "Unable to stop 
turning." Their last flight in 1904 was on December 
7 and the problem of turning was still unsolved. 

The difficulties the Wrights encountered in turns 
were only partly due to the spiral instability. They 
believed later (Reference 1, footnote, pp. 469-71) 
that the control system was a serious cause as well. 
In all of the flights referred to above, the wing warping 
and rudder deflection were interconnected as in the 

16. Wright airplane, 13 August 1904, showing wing rigged with anhedral. (From The Papers of Wilbur and Orville 
Wright, Marvin W. McFarland, editor, New York, 1953) 


F.E.C. Culick and Henry R. Jex 

17. Wright airplane, 16 November 1904, showing wing rigged without anhedral. (From The Papers of Wilbur and 
Owille Wright, Marvin W. McFarland, editor, New York, 1953) 

1903 Flyer. They recognized that this restricted the 
control they had and finally in 1905 decided to operate 
the control independently. 

At the beginning of the tests in 1905 (late August) 
the wings were rigged with a small amount of anhedral 
which was later removed. Together with independent 
control of yaw and roll, this gave the Wrights an 
airplane they could turn controllably at speed and 
altitude. They then discovered the last problem they 
had to solve to have a practical airplane: stalling in a 
turn. Between September 28 when they first flew in 
1905 with independent warp and rudder, and October 
5 when they flew for 38 minutes, the Wrights learned 
how to recover from a stall. Wilbur's description in 
his summary of the experiments in 1905 (Reference 
1, pp. 519-21) is a superb statement of the problem 
and its solution: 

The trouble was really due to the fact that in 
circling, the machine has to carry the load resulting 
from centrifugal force, in addition to its own weight, 
since the actual pressure that the air must sustain is 
that due to the resultant of the two forces . . . 
When we had discovered the real nature of the 
trouble, and knew that it could always be remedied 
by tilting the machine forward a little, so that its 
flying speed would be restored, we felt that we were 
ready to place flying machines on the market. 

What a magnificent achievement! In the seven days 
from September 28 to October 5, 1905, the Wright 
brothers solved their last serious problem and had a 
practical airplane. They didn't fly again until 1908, 
but that's a different part of the story. 

Lateral Aerodynamics: Warping Effectiveness 

One of the major purposes of the wind tunnel tests 
with the Vb scale covered model was to investigate 

the quantitative aspects of wing warping. This method 
of lateral control was original with the Wrights and 
after their first flights in 1908 it was quite widely 
adopted. 6 But within five years it had been almost 
entirely discarded in favor of ailerons. Hence no wind 
tunnel data had been taken for the performance of 
warping. It is an important matter of historical docu- 
mentation to establish quantitatively how this method 
of control worked. Some of the results of the GALCIT 
tests are summarized in Figure 18. 

The top portion of Figure 18 shows the effects of 
warping the wing with no rudder deflection. Data are 
plotted for no warp (open circles) and maximum warp 
(open triangles). As noted earlier in connection with 
Figures 14 and 15, a non-zero roll moment exists with 
no warp deflection because the starboard wing is 
longer than the port wing. The roll moment produced 
is slightly dependent on a, the angle of attack. 
However, the adverse yaw moment accompanying 
the warp is strongly dependent upon a. It is adverse 
yaw in the sense that a right turn produces a yaw 
moment tending to turn the nose to the left. 

Wilbur made his fundamental discovery of adverse 
yaw during his flights in 1901. He noted in his diary 
on August 15, "Upturned wing seems to fall behind, 
but at first rises." Then in a letter to Chanute on 
August 11, he wrote, "The last week was without 
very great results though we proved that our machine 
does not turn (i.e., circle) toward the lowest wing 
under all circumstances, a very unlooked for result 
and one which completely upsets our theories as the 
causes which produce the turning to right or left." 
These are the first observations of adverse yaw. They 
could only be made by someone who understood 
something of aerodynamics and flight mechanics but 

Aerodynamics, Stability, and Control 


a) Wing Warp Alone 


No warp control 


Maximum warp 

(» ± 5°eachtip) 

A^ Wing Warp * L inked Rudder 

18. Data for the lateral effectiveness (Vt, scale model). 

especially was trying to learn to fly and was a keen 

Adverse yaw arises in the following way. In order 
to turn, as the Wrights understood from the beginning 
of their work, it is necessary to generate a component 
of force towards the center of the turn. This is best 
accomplished by tilting the lift force on the wing, 
which is done by banking the entire aircraft. A bank 
is produced by applying a roll moment, generated by 
increasing the lift on one wing and reducing the lift 
on the other. When that happens, whether by wing 
warping or by using ailerons, the drag is increased on 
the wing carrying more lift and reduced on the other. 
The differential drag acts as a yaw moment tending 
to swing the nose of the aircraft in the direction 
opposite to that of the desired turn — hence the name 
adverse yaw. It inevitably accompanies any turning 
maneuver. Although adverse yaw is low at higher 
flight speeds, and can be reduced with clever design 
of the lateral control system, what is really required 
is control in yaw, and that is why a vertical control 
surface or rudder must be installed. 

The most fundamental aspect of the Wrights' in- 
vention of the airplane was the idea of the need for 
control of both roll and yaw motions. It is the 
foundation of their basic patent submitted in 1902 
and granted in 1906. Wilbur had discovered the 
problem of adverse yaw in 1901. Their first glider in 

1902 had a fixed vertical tail which, with anhedral, 
gave flying characteristics which they considered to 
be the most difficult of all their aircraft. They quickly 
installed a moveable tail which of course gave them 
the necessary control in yaw. 

Warp and rudder deflections were interconnected 
in the 1902 glider and in the 1903 airplane. Although 
the controls were later made independent, intercon- 
nection was a fortunate choice for the 1903 machine, 
as we shall see in the following section. The data 
plotted in the lower portion of Figure 18 shows how 
simultaneous deflection of the rudder with warping 
compensates for adverse yaw. The curve labeled 8 r 
= 12.5° crosses the axis, indicating zero yaw moment, 
at 8 r = 4°. For the covered model (see Figure 12) 
this is nearly the angle of attack for the cruise 
condition. Thus for this speed only, this combination 
of warp and rudder deflection will produce a roll 
moment with no adverse yaw, which allows entry to 
a banked turn with no sideslip — i.e., a more coordi- 
nated turn. 7 By disconnecting the warp and rudder 
controls in their 1905 airplane, and installing both 
controls on a single stick, the Wrights were then able 
to execute coordinated turns over a range of airspeeds, 
in a convenient fashion. 

Summary of Wind Tunnel Tests of the 1903 Flyer 

The results of these wind tunnel tests have greatly 
increased our understanding of the flying character- 
istics of the 1903 Flyer. It appears that the data are 
reasonable and agree well with predictions based on 
modern aerodynamic theory. 

According to these data, the trimmed flight condi- 
tion of the aircraft is near the optimum, being at a 
value of lift coefficient slightly less than that for 
maximum lift/drag ratio. This provided ample margin 
below stall of the aircraft, a primary consideration 
particularly in view of Lilienthal's fatal crash. 

The canard gave sufficient power in pitch to control 
the unstable motions, and the vertical tail was ade- 
quate to contol yaw. The combination of wing warp 
for roll control and a link to remove the associated 
adverse yaw provided powerful lateral control for 
banking the airplane and for coping with gusts. No 
contemporary aircraft had control even approximating 
that of the 1903 Flyer until after the Wrights publicly 
flew their improved airplane in 1908. 

Dynamical Stability and Control 

Our discussion of the wind tunnel data has verified 
and clarified most of the important static characteristics 
of the 1903 Flyer — static stability and control effec- 
tiveness. With our data, and estimates of a few 
quantities, we are able to describe quite accurately 
the dynamics of the airplane, in quantitative terms 


F.E.C. Culick and Henry R. Jex 

not available to the Wrights. 

Because the Flyer logged a total flight time of only 
1 minute 58 seconds, the flight characteristics and 
handling qualities of the airplane were never fully 
tested. That it was flyable was of course demon- 
strated — under severely gusty conditions. In this sec- 
tion we try to convey some idea of how the airplane 
probably behaved, by examining two elementary 
transient motions of pitching and turning. 

First a few general remarks on unsteady or dynam- 
ical motions of aircraft. We assume that the airplane 
has a plane of symmetry containing the longitudinal 
and vertical axes. 8 It is then a general theoretical 
consequence of the equations of motion that if the 
disturbances away from steady flight are not too large, 
then the unsteady motions can be split into two parts: 
purely longitudinal motions involve changes of the 
forward speed, pitch attitude, and vertical speed, or 
angle of attack. The lateral motions are out of the 
plane of symmetry, comprising roll, yaw, and sideways 
translational motion or sideslip. 

The practical consequence of this general splitting 
or uncoupling of the motions is that, for example, 
movement of the pitch control (elevator or canard), 
or a purely vertical gust, will not generate lateral 
motions out of the plane of symmetry, and conversely. 
This is the reason why we can rigorously treat the 
pitch dynamics separately from the lateral dynamics. 
It is a good approximation to actual motions. 

Dynamics of Pitching Motions 

We have already established that the Wright Flyer 
was statically unstable in pitch. That means that if it 
is even slightly disturbed from a condition of steady 
flight, there is no tendency to restore the initial steady 
motion. Thus if the pilot does nothing, the airplane 
will exhibit a divergent nose-up or nose-down depar- 

Figure 19 shows the results of a calculation. Suppose 
that in level cruise flight 9 the pilot suddenly deflects 
the canard nose-up one degree and immediately 
returns it to its previous setting. The same input can 
be imaged due to an infinitesimally short vertical gust 
having speed roughly 3 A foot per second, a mild gust. 
This pulse input is represented in Figure 19a. The 
remaining four parts of the figure clearly show the 
subsequent divergent motions in angle of attack, pitch 
(nose up), airspeed (decreasing), and altitude (increas- 
ing). In approximately one-half second the amplitude 
of the motion doubles. Thus, if the angle of pitch is, 
say five degrees at some time after the canard has 
been pulsed, then the pitch angle is already ten 
degrees only one-half second later. 

The airplane alone is obviously very unstable both 
statically and dynamically. However, it can be con- 
trolled by a skilled pilot — the practical consequence 

is that the combination of airplane plus pilot is a 
dynamically stable system. It is entirely analogous to 
the manner in which a statically unstable bicycle with 
a trained rider is stabilized. So far as reaction time is 
concerned, stabilizing the 1903 Flyer is roughly equiv- 
alent to balancing a yardstick vertically on one's finger! 
Practice is required — the Wrights had lots of that. 
Here, to demonstrate the idea, we assume that in 
response to a disturbance the pilot tries to maintain 
level flight with a simple strategy. The pilot can see 
the horizon and he knows where some horizontal 
reference line on the canard should lie with respect 




of Attack 



Pulsed for 
I deg-sec 

-5 J 

-5 J 

Pitch Angle 
of Skids 


-5 J 



U 30 

-5 J 

Stalls at 
2 deg 

23 mph 

Time (sec) 

19. Open loop time response in pitch; one degree 
pulsed canard deflection. 

Aerodynamics, Stability, and Control 


to the horizon in level flight. Then to restore level 
flight, the pilot deflects the canard by an amount 
which is proportional to the error between the actual 
location of the reference line and its desired position 
in level flight. Thus, the canard deflection is propor- 
tional to the pitch error; the constant of proportionality 
is called the pilot's "gain." 

The airplane and pilot, with the assumed propor- 
tional control, constitute a feedback system. We 
interpret its behavior in a root locus diagram, sketched 
in Figure 20. It is not appropriate here to discuss the 
theory of this diagram; we shall only explain briefly 
its meaning and the implications of the results. 

Return to Figure 19, at the top, the feedback 
system comprising the airframe plus pilot is repre- 
sented as a block diagram. The equation labeled 
"open loop" is used to calculate the response of pitch 
angle, 8, to a sinusoidal variation of canard deflection 
with maximum excursion + 5 e (nose up) and — 8 e 
(nose down). With suitable operations, this formula 
can be extended to compute the response in pitch to 
any variation of canard deflection; that is how the 

results shown in Figure 19 were found. These results 
follow from the complete linearized equations for 
longitudinal motions; their derivations will not be 
described here. Reference 14 contains a thorough 
coverage of the theory. The paper by Professor 
Hooven in this collection shows how to compute the 
real-time response using a digital computer as a 

The denominator of the open loop response is 
shown as the product of three factors, one labeled 
"phugoid" and two together identified as "short 
period." It is helpful in explaining Figure 20 to remark 
briefly on the origin of these terms. 

We have already noted that under quite general 
conditions, the longitudinal dynamics can rigorously 
be treated separately from lateral motions. For most 
aircraft, there are two fundamental modes of longi- 
tudinal motion, called the short period and phugoid 
oscillations. The phugoid was discovered, analyzed, 
and named in a remarkable work by F.W. Lanchester 
in the mid 1890s (Reference 15) based on his obser- 
vations of the flights of model aircraft. 10 This is a 

Desired Pitch 
Pitch Angle 

Angle Error 





Pilot Airframe 





Feedback (visual) 


at K p = 4 

Open Loop- 


M Se I/T0, l/T0 2 
ll.0(s + .5)(s + 3.0) 

o~ ~ y p* y 8* = k p t~? — : — r. — : t\ ; k d opt =4.0 

9e w ° e M [s 2 +2(.30)(l.2)s + l.2 2 ](s-l.7)(s + 7.0) P 

l/Tsp, l/T S p 2 


Phugoid Mode Short Period Modes 

Closed Loop : 

(s + .5)(s + 3.0) 

C [s 2 -t-2(.IO)(5.5)s + 5.5 2 ](s + .33)(s + 4.6) 
I Closed loop poles for K p =4.0deg canard/deg pitch error 

I -6 

J— t 


I -4 

'sp 2 





I Hz 





20. Root locus diagram for pitching motions; proportional control law for pilot-closure (pitch angle error). 


F.E.C. Culkk and Henry R. Jex 

relatively slow undulating motion whose behavior is 
dominated by the interchange of kinetic energy of 
forward motion and potential energy of vertical mo- 
tion. The angle of incidence remains nearly constant 
while the pitch angle changes, being horizontal near 
the maxima and minima of the undulations. It is the 
phugoid mode which causes difficulties in trimming 
aircraft when changes of pitch attitude are made. 

The second fundamental mode of motion, the short 
period oscillation, normally has frequency much higher 
than that of the phugoid mode. Now the aircraft 
behaves as an oscillator or weathervane in pitch, the 
mass being the moment of inertia in pitch and the 
"spring" being proportional to the static stability in 
pitch, the static margin. The forward speed remains 
nearly constant and the nose bobs up and down with 
the angle of incidence approximately equal to the 
angle of pitch. Because the tail (or canard) also moves 
up and down with the periodic motions, there is 
considerable damping of the motion. It is the short 
period oscillation which usually tends to be most 
easily excited by sharp gusts and turbulence. 

Now back to the 1903 Wright Flyer. In the context 
of aircraft dynamics, this is distinctly not a conven- 
tional machine, which makes its study particularly 
interesting. First we find that, because the airplane 
is statically unstable in pitch, the usual short period 
oscillation doesn't exist. It degenerates to two simpler 
fundamental motions, one of which decays with time 
and the other of which diverges following a disturb- 
ance. The latter is responsible for the behavior shown 
in Figure 19. The phugoid is lightly damped, as 
normally true, and has a period of about six seconds. 
A typical general aviation aircraft will have a period 
of say 30-40 seconds for the phugoid and less than 1 
second for the short period oscillation. Hence what 
we call here the "phugoid" is really something be- 
tween the conventional phugoid and short period 

The coordinates in Figure 20 are the angular fre- 
quency co in radians plotted vertically, and decay or 
growth constant, 1/T plotted horizontally. The period 
of motion is 2tt/co and the amplitude of motion varies 
as exp (t/T). Thus, if 1/T is negative — i.e., lies on 
the left side of the diagram, the motion decays, 
proportional to exp ( — t/T) and after t = T seconds 
the amplitude is reduced by a factor of about .37. 
The crosses in Figure 20 denote the roots of the 
denominator of the formula for 0/8 e and represent the 
natural motions when the pilot does nothing — the 
canard surface remains fixed. These points are labeled 
(Dp, denoting phugoid, and 1/Tsp,, 1/Tsp,, denoting 
the degenerate short period. Note as remarked above 
that one of the latter two lies to the right of the 
vertical axis, representing a divergent motion, and 
one lies to the left. 

Now suppose the pilot acts as described earlier, 
and continually deflects the canard in opposition to 
the perceived pitch deviation to maintain a desired 
pitch attitude — the "loop is closed." The fundamental 
motions of the complete system, aircraft plus pilot, 
must clearly be different from those for the "open 
loop," or aircraft alone. A different formula for 8/8 c 
is found and the roots of its denominator are different 
from those plotted as the crosses. In particular, the 
values of the roots depend on the gain, K p , of the 
pilot — how much he deflects the canard for a unit 
perceived error. As K p is changed, each root traces a 
locus starting at the open loop cross, and hence the 
name "root locus diagram." 

The filled squares in Figure 20 represent the roots 
when K p = 4, meaning that the pilot deflects the canard 
by four degrees for every degree of error he sees. 
Both roots on the horizontal axis now represent stable 
motions which always decay. The root representing 
the oscillation has now moved to higher frequency 
and is still lightly damped. This frequency, roughly 
0.9 Hertz, the period being about 1.1 seconds, is in 
the range for which pilot-induced oscillations will 
occur. They were likely a problem for the 1903 Flyer, 
as shown by photographs in which the canard is 
deflected fully up or down. Motion pictures of the 
Wright aircraft in 1909 confirm these unstable ten- 

Figure 21 is a sketch of the time response for a one 
degree deflection of the canard, corresponding to the 

deg elevator 

Kfl. = 4.0 . .. . 

P deg pitch error 






4 6 

Time (sec) 

21. Closed loop time response in pitch; one degree 
pulsed canard deflection. 

Aerodynamics, Stability, and Control 


case shown in Figure 19, but now the pilot exercises 
proportional control (K p = 4). Both the horizontal speed 
and the height are successfully maintained constant, 
but the nose bobs up and down at about 1.1 cycles 
per second; after about two and one half cycles the 
amplitude is reduced by half. Thus we have found 
that even though the airplane alone is seriously 
unstable in pitch, it is controllable by a reasonably 
skilled pilot. 

This behavior more closely resembles the short 
period motion than it does the phugoid. As we noted 
above, the lightly damped oscillation of the airplane 
alone really cannot be called a phugoid and we have 
here further support for this view. The origin of this 
unusual behavior is of course the unorthodox combi- 
nation of aerodynamic characteristics, including the 
unstable configuration, and its inertial properties. 
Having a wingspan of 40 feet, the 1903 Flyer was 
quite large, but its wing loading was only 1.5 pounds 
per square foot, which places it in the class we now 
call ultralights. One important consequence of the 
low wing loading is that the mass of air which must 
be moved in accelerated motions — the virtual mass 
and virtual inertia — is a significant fraction of the mass 
of the airplane; here about 20 percent. This has been 
accounted for in the results shown, and explains part 
of the peculiar behavior. 

Approximate values of the virtual inertia have been 
used in the results given here. The calculations are 
being refined for a biplane cell having finite aspect 
ratio. However there is no doubt that the oscillatory 
motion shown in Figure 21 is real. Films of the 
Wrights flying their improved aircraft in 1909 show 
clearly exactly this kind of continuously oscillating 
pitch control at about the same frequency. 

Dynamics of Lateral Motions 

The Wright brothers were the first to understand the 
correct method for turning an airplane. Lilienthal and 
other glider pilots he inspired were largely content to 
maintain lateral equilibrium by building wings with 
dihedral, and shift their weight as required during 
flight. Contemporary experimenters with early pow- 
ered aircraft, such as Gabriel Voisin in France, tried 
to skid around turns by deflecting the rudder. Only 
the Wrights realized that good roll control is essential 
for turn entries and exits. They devoted a large part 
of their flight test program to the problem of turning; 
only after they were satisfied with their solution did 
they set out to sell their invention. We have discussed 
the main features of their system for control of roll 
and yaw of the 1903 Flyer. Now let us see how it 
actually performed in flight. 

According to discussion in the preceding section, 
one can treat the lateral motions independently of 
pitching motions. Before analyzing the particular be- 

havior of the Flyer, it is helpful to consider some 
elementary characteristics of a turning maneuver. 
Imagine an airplane in steady level flight, and suppose 
that a means for applying a roll moment is available, 
by deflection of ailerons, or by wing warping. A fixed 
value of deflection or warp generates a constant roll 
moment. If a constant roll moment is suddenly ap- 
plied, the airplane is first accelerated in roll, but soon 
settles down to a constant roll rate, so the bank angle 
increases linearly in time. 11 The rate is constant 
because the moment due to the distorted wings is 
compensated by the damping in roll, a moment 
opposing the movements of the large surface areas 
normal to themselves. Figure 22 shows this behavior 
for the 1903 Flyer, the lateral response for an impulsive 
warp deflection, two degrees of warp held for one- 
half second, with no rudder deflection. The unstable 
nature of the motion is clearly shown by the rapid 
divergence of roll and sideslip angles. Note that owing 
to adverse yaw, the heading rate is initially in the 


Rate ^ 


-10 J 

10 -i 

S w 

-10 J 

Impulse = 1.0 deg-sec 

J— L 


Time (sec) 

22. Open loop time response for roll angle; ten 
degrees of wing warp. 


F.E.C. Culick and Henry R. J ex 

direction opposite to that desired. 

Evidently, to execute a turn with a fixed bank 
angle, the roll moment must first be turned on and 
then removed. Simultaneously, the rudder must be 
used in such a fashion as to compensate adverse yaw 
and reduce the sideslip to zero. Considerable practice 
is required to perform smooth turns. 

Analysis of the turn may be carried out using the 
methods described above. We require that, beginning 
from steady level flight, the pilot actuate the controls 
in such a manner as to roll the airplane into a constant 
angle of bank. The root locus diagram in Figure 23 
has been constructed for this situation. Below the 
block diagram is the equation labeled open loop 
response, a formula for the response of roll angle to 
wing warp, 8 M . The crosses in the diagram again 
represent the roots of the denominator. One lies to 
the right of the vertical axis, and corresponds to the 

unstable spiral mode described earlier. If the wings 
are impulsively warped, and returned to their undis- 
torted state, or if the airplane is exposed to a short 
vertical gust unsymmetrical about the centerline, a 
divergent spiral motion will develop, as previously 

Another root lies far to the left; this is labeled "roll 
subsidence" and arises from the heavy damping of 
roll motions. The third root, w DR , presents a damped 
oscillation, the subscript DR standing for "Dutch 
roll." 12 This is primarily an oscillation in yaw angle, 
a mode due to the action of the vertical tail as a 
weathervane. This induces oscillatory motions in both 
roll and sideslip. Damping of the motion is provided 
mainly by the vertical tail and drag of the wings and 
struts, due to the differential airspeeds accompanying 
yaw rates. 
These three modes — the spiral mode, the roll subsid- 

Bank Percieved 

Command Error 





Linkage |_ 











Open Loop (pure warp) ■ 



5.9 [.005,1.3] 

P 'YSw = K P* (-3)(3.8)[.6,l.3] 
'/T s l/T R £ DR u) DH 

Open Loop (warp + linked rudder) : 

£ - K 5 - 8 l* 7 - 1 -2] 

c/> e " P ' (-.3) (3.8) [.6, 1.3] 


Closed Loop » 



.25 , 





L l *" 





u) D 


* . 



— r - 






a) Warp Alone 

b) Linked Rudder (S r =-2.5 S w ) 

23. Root locus diagram for lateral motions; proportion control law for pilot closure (roll error). 

Aerodynamics, Stability, and Control 


ence, and the Dutch roll — are the natural lateral 
motions of all aircraft. In this respect, the lateral 
behavior of the 1903 Flyer is generically the same as 
conventional aircraft, which normally can be charac- 
terized by the same lateral modes. However, the spiral 
mode is unusually unstable, the amplitude doubling 
in about 2.5 seconds. This rapid growth is due largely 
to the anhedral as we discussed earlier. Partly because 
of the low speed and partly because of the low 
directional stability compared with the large yaw 
inertia, the period of the Dutch roll oscillation is 
relatively long, roughly 4.5 seconds. It is not heavily 
damped due to the relatively small vertical tail and 
hence small damping in yaw. 

Now consider a simple model of a turn maneuver. 
Suppose that the pilot wishes to obtain a bank angle 
equal to ten degrees, which he observes as the angle 
between the horizon and the canard reference line. 
As a control law we assume the pilot operates the 
warp control by an amount proportional to the error, 

the difference between the desired bank angle (10 
degrees) and that actually observed; the constant of 
proportionality is the gain, Kp. Two cases are treated: 
pure warp, with no deflection of the rudder; and 
interconnected warp/rudder. The second corresponds 
to the control system in the 1903 Flyer; the drawings 
obtained from the Smithsonian Institution imply that 
the rudder is deflected —2.5 degrees for each degree 
of warp deflection. As for the longitudinal motions, 
the locus of roots can be calculated for the two cases, 
shown in Figure 23. For increasing gain, the roots 
corresponding to the spiral mode and roll subsidence 
move towards each other on the horizontal axis and 
then depart vertically, representing the formation of 
a heavily damped "spiral roll" mode whose dynamics 
characterize the major portion of the response in roll. 
More interesting is the dependence of the Dutch 
roll "nuisance" oscillation on the gain. For the case 
of pure warp, this becomes marginally damped for a 
reasonable value of the gain, one degree of warp for 



10 -i 

S r 

-10 J 

H 1 1 1 1 1 1 1 1 1 1 1 P 1 1 1 1 1 1 1 

Turn Rate =10 deg/sec 

-i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — 

«* * ' — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i 

H *SL 1 1 1 F 1 1 1 1 1 P 1 1 1 1 1 1 1 

K p = 1.0 

S r =-2.5S W K p = I.O 

' ' i i i i i i | i i i i i i i i i i 
5 10 15 20 

Time (sec) 

*-A\ — i — I — i — i — i — i — i — i — i — i — i — | — i — i — i — »- 

5 10 15 

Time (sec) 

a) Warp Alone b) Rudder Linked to Warp 

24. Closed loop time response of lateral motions; ten degrees of wing warp. 



F.E.C. Culick and Henry R. J ex 

each degree of perceived error. The time history for 
this motion is shown in Figure 24a. Large oscillations 
of both bank angle and sideslip make this a wallowing 
motion nearly impossible to control and wholly un- 
satisfactory for practical flying. It is mainly due to the 
combination of anhedral and uncompensated adverse 

When the rudder deflection is linked to the warping, 
thereby canceling the adverse law, the result is a 
turning maneuver which is quite acceptable. The 
closed loop damping is now much higher — the filled 
square in Figure 23 lies well to the left of the vertical 
axis. The much improved response in time appears 
in Figure 24b. Now the bank angle tends to a constant 
value, albeit not equal to the desired value ( 10 degrees) 
within the time scale shown. There is a fairly large 
angle of sideslip, so it is a sloppy uncoordinated turn, 
but surely possible. Thus the interconnection of the 
warp and rudder is an essential feature of the 1903 

As the Wrights discovered in 1905, satisfactory 
control is achieved only by warp and rudder coordi- 
nation more complicated than proportional intercon- 
nection. It has often been stated, incorrectly, that the 
Wrights abandoned their interconnected warp and 
rudder. In their 1908 airplane, with the pilot sitting 
upright, they put both rudder and warp controls on a 
single stick. Lateral hand motion caused warp, while 
fore-and-aft motion deflected the rudder. Conse- 
quently any desired proportion of warp and rudder 
could be produced by operating the stick in a suitable 
diagonal path. Far from abandoning warp/rudder in- 
terconnection, the Wright brothers ingeniously pro- 
vided a ratio instantly adjustable according to the trim 
speed or angle of attack. The data discussed earlier 
(Figure 12) suggest the need for this flexible control. 

Further evidence that the Wrights had a most 
advanced understanding of aircraft control appears in 
a short French monograph published in 1909 (Ref- 
erence 16). M. Pol Ravigneaux, evidently instructed 
by the Wrights, gave a detailed analysis of the stick 
movements required to accomplish various lateral 
motions. A few remarks taken from the discussion of 
his explanation illustrate the point. 

Any movement of the lever L from right to left, or 
vice versa. . .produces warping which is inverse at 
the tips of the two lifting surfaces. Any motion of the 
lever L forward or backward causes. . .a "rotation" 
of the vertical directional rudder. . .By actuating this 
lever obliquely, one will obtain simultaneously warping 
and movement of the rudder. 

We know that warping which causes a [left] bank 
causes simultaneously a [right] turn. . . 

To prepare and make a turn to the left: 1) bank to 
the left by warping the wings and beginning to turn; 
2) straighten out the warped surfaces so as not to 
continue the banking and smartly turn and rudder to 
the left. To finish the turn: 3) straighten the rudder; 

4) level the machine by reversing the previous warp; 

5) return the wing surface and the rudder to their 
neutral states. 

The author then notes that in practice, the steps 
describing initiation and completion of a turn overlap, 
so the use of warp and rudder deflections are executed 
more continuously to produce smoother turns. 

No contemporaries of the Wrights possessed such 
a thorough appreciation of the details of turn coordi- 
nation. Our analysis of the dynamics verify the sound- 
ness of the Wrights' concepts for lateral control. The 
results give us even more respect for their ability to 
accomplish nearly perfect turns. 

Concluding Remarks 

In 1903, the Wrights understood well the subjects of 
structures, performance, and control. Their structural 
design is discussed elsewhere in this collection. Their 
craftsmanship far exceeded that of their contempor- 
aries. Performance is essentially a matter of balancing 
forces: weight, lift, drag, and thrust. The theory 
required is minimal. But it seems clear from analysis 
of our wind tunnel data, combined with the docu- 
mented characteristics of their engine and the 1903 
airplane, that the Wrights must have paid much 
attention to this problem. It is not likely accidental 
that the geometrical incidence of the wing was set at 
the angle of incidence for cruise flight. Nor was it a 
matter of luck that the cruise condition gave them a 
good margin below stall of the wing. 

They had learned from Lilienthal that to design a 
successful airplane they also had to learn to fly. What 
they added to that lesson was control, unquestionably 
their greatest contribution. From the beginning of 
their work they knew that they had to control rolling 
and not just pitching as their contemporaries had 
emphasized. Later they discovered that they also had 
to control yaw motions. That eventually made the 
1903 Flyer manageable. 

We have used recent wind tunnel data and modern 
theory of stability and control to confirm the Wrights' 
unparalleled understanding of aircraft control. Solu- 
tion of the problem of turning was their supreme 
achievement in flight dynamics and gave them a 
marketable airplane. Their success required appreci- 
ation of aerodynamics and invention of a simple means 
for the pilot to exercise lateral control with coordinated 
wing warping and rudder deflection. 

There was much the Wrights did not understand 
well, mainly subjects which were not clarified until 
many years later. Perhaps the greatest gap in their 
knowledge was the theory of rotational motions. 
Without that they could not formulate precise ideas 
of stability in contrast to equilibrium. 

Their 1903 Flyer was severely unstable statically, 

Aerodynamics, Stability, and Control 


and barely controllable by modern standards of pilot- 
ing. They detected the most serious difficulties during 
flight tests in 1904 and 1905, but could correct them 
only by trial-and-error: they had no guiding theory. 
For example, they had deliberately used negatively 
arched wings to combat the tendency for lateral gusts 
to force them into the hill while gliding. Our analysis 
of the dynamics has shown that as a result of the 
negative dihedral, the spiral mode was so strongly 
unstable as to be marginally controllable. The Wrights 
spent nearly a year at Huffman Prairie before they 
removed the negative dihedral in 1904. They had 
been treating the instability as a problem of lateral 
control, but it was in fact a problem of lateral dynamics. 

The Wrights' emphasis on control unquestionably 
flowed from their experience with bicycles. They 
knew that their airplane need not be inherently stable 
to be flyable. Their creation of the first practical 
aircraft proved their principles. 

The achievements of the Wright brothers appear 
more remarkable the deeper we understand their 
technical work. Their own thorough documentation 
in letters and diaries makes it possible to interpret 
their work in the context of modern aeronautics. It is 
astonishing how rarely they strayed from the system- 
atic path to success. What they could not solve with 
theory and analysis they figured out with careful 
testing and observations. The standards they set as 
aeronautical engineers remain unsurpassed. 

F.E.C. Culick received his Ph.D. in Aeronautics and 
Astronautics with a minor in Physics from MIT in 1961. As 
Professor of Applied Physics in Jet Propulsion at Cal Tech 
he teaches courses in both fields and currently lectures on 
Applied Aerodynamics and Flight Mechanics. 

Professor Culick is a member of the American Physical 
Society and the American Institute of Aeronautics and 
Astronautics. He has published numerous papers and is 
listed in Who's Who in America and American Men and Women 
of Science. 

Henry R. Jex is Principal Research Engineer at Systems 
Technology, Inc., in Hawthorne, California. He holds a 
B.S. in Aeronautical Engineering from MIT and a M.S. in 
Aeronautical Engineering from Cal Tech. Since 1953, Mr. 
Jex's professional experience has been devoted to all aspects 
of aerodynamic design, flight control integration, and man- 
machine dynamics, including the conception, develop- 
ment, and testing of aircraft employing minimal flight con- 
trol systems, tailored airframe characteristics, and novel 
control surface arrangements. He did the control system 
design and analysis for the successful Gossamer-Condor 
and -Albatross human powered aircraft. 

Mr. Jex has been an invited lecturer or delegate to several 
government, academic, or international organizations and 
has published numerous technical papers and monographs. 
He is a member of Tau Beta Pi, Gamma Alpha Rho, the 
American Institute of Aeronautics and Astronautics, Low 
Speed Aerodynamics Research Association, and the Human 
Factors Society. 


1. In a letter written on December 22, 1903, Bishop 
Milton Wright, father of the Wright brothers, referred to 
their aircraft as the "Flyer" (Reference 1). This seems to 
be the earliest use of the name. Whether or not Bishop 
Wright intended to give the aircraft an "official" name is, 
we suggest, immaterial. He used it, it's a good name, and 
arguments as to its correctness, in some sense, seem 
pointless. We subscribe to Gibbs-Smith's usage 
(Reference 2). 

2. It is a remarkable consequence of progress that some 
of the most advanced aircraft designs are based on 
unstable configurations, stabilized with automatic flight 
control systems. These are called "control-configured 
vehicles." The Wright brothers deserve recognition as the 
first proponents of this "modern" approach to design. In a 
further twist of fate, these control-configured vehicles are 
plagued by many of problems discovered by the Wrights! 

3. We must hedge a bit. The right wing of the 1903 
Flyer was about four inches longer than the left, to 
compensate for the weight of the engine, which was 
mounted to the right of the pilot. This is clear evidence 
of careful design, and an indication that the Wrights 
understood some of the need to balance moments as well 
as forces. 

4. This convention is historical and originated with early 
plots of wind tunnel data prepared in Great Britain. With 
this convention, the moment curves for stable aircraft fall 
to the right of the diagram three plots — the drag polar, 
the lift curve, and the moment curve — could be placed 
side-by-side on one sheet of paper. Theorists, on the 
other hand, often do not follow this convention! 

5. Plates 84 and 86 of Reference 1. 

6. The Wrights used a Pratt truss between the lower 
wings; the vertical struts carry compressive loads and 
diagonal wires carry loads in tension. This design 
provided a rigid, arched "beam" as the forward section of 
the biplane. The center portion of the biplane was also 
rigidly trussed at the aft spars. But the outboard 40 
percent of the aft spars were trussed by a set of wires to 
permit controlled warping. When the trailing edges of one 
pair of tips are twisted up, the trailing edges on the 
opposite side twist down. Clever structural design is 
necessary to reduce the wings' resistance to warping so 
that the control forces are not too large. Those working 
on the AIAA Wright Flyer project have great respect for 
the Wrights' ingenious solution to this problem. 

7. This conclusion is not wholly correct because our 
discussion is oversimplified and incomplete. We have 
ignored the effects of the sideforce generated by rudder 

8. The assumption is only slightly strained because of the 
deliberate asymmetry mentioned earlier. This has very 
small effects on the results. 

9. Because the airplane is unstable this condition can in 
reality exist only for a brief time. For calculations we can 
ignore that practical problem and assume that we start 
from the desired state of nice level flight. 

10. Lanchester chose the term phugoid based on Latin 
and Greek roots meaning "to fly." He mistakenly 
selected roots meaning to fly in the sense of to flee — as in 
"fugitive." Lanchester's aerodynamics was much superior 
to his etymology. 

11. Note that in contrast, fot a stable aircraft, fixed 
deflection of the elevator, which produces a constant 


F.E.C. Culick and Henry R. Jex 

change of the pitching moment, causes a constant change 
of pitch angle (or angle of incidence), not a constant pitch 
rate. This is different from roll motion because the 
pitching moment due to the elevator is compensated by a 
change in the pitching moment due to the lift of the 
wing. If the aircraft is unstable in pitch, as the 1903 Flyer 
was, the two contributions to the pitching moment act 
together and the pitch attitude of the airplane diverges. 

12. The origin of the term "Dutch roll" is obscure. The 
eminent aeronautical scientist Theodore von Karman once 
explained that it was a contraction of the naval jargon 
"Dutchman's roll," alluding to the motion of round- 
bottomed Dutch ships in the North Sea, or of round- 
bottomed Dutch sailors ashore. Take your pick. 


1. McFarland, M., ed. The Papers of Wilbur and Orville 
Wright. New York: McGraw-Hill Book Co., 1953. 

2. Gibbs-Smith Charles H. The Invention of the Aeroplane. 
New York: Taplinger Publishing Co., 1965. 

3. Wenham, F.H. and J. Browning. Aeronautical Society 
of Great Britain, Annual Report, 1871. 

4. Wells, Albert C. "An Investigation of Wind Pressures 
upon Surfaces." Thesis. MIT, 1896. 

5. Lilienthal, Otto. Birdflights as the Basis of Aviation. 
Translated by A.W. Isenthal. London: Longmans, Green 
and Co., 1911. 

6. Smeaton, John "An Experimental Inquiry Concerning 
the Natural Powers of Water and Wind to Turn Mills and 
Other Machines Depending on a Circular Motion." 1752. 

7. Langley, Samuel P. Experiments in Aerodynamics. 
Smithsonian Institution Publication 801, 1891. 

8. Penaud, A. "Aeroplane Automoteur; Equilibre 
Automatique." L'AeronauteWo\. 5. 1872, pp. 2-9. 

9. Bryan, G.H., and W.E. Williams. "The Longitudinal 
Stabilitv of Gliders." Proc. Roy. Soc. of London, Vol. 73. 
1904, pp. 100-16. 

10. Turnbull, W.R. "Researcher on the Forms and 
Stability of Aeroplanes." Physical Review, Vol. 24, No. 3. 
March 1907, pp. 285-302. 

11. Bettes, W.H., and F.E.C. Culick. "Report on Wind 
Tunnel Tests of a Vb — Scale Model of the 1903 Wright 
Flyer Airplane." Guggenheim Aeronautical Laboratory, 
California Institute of Technology, GALCIT Report 
1034. 1982. 

12. "1903 Wright Flyer Ve Scale Model Wind Tunnel 
Aerodynamic Data." 1982. 

13. Root, Amos. "My Flving Machine Storv." Gleaning in 
Bee Culture. Vol. 33. January 1905. pp. 36-39, 48. 
Reprinted with title "The First Eye-witness Account of a 
Powered Airplane Flight (1904)" in The Aeroplane; An 
Historical Survey of Its Origins and Development, by Charles 
H. Gibbs-Smith. London: Her Majesty's Stationers 
Office, 1960, pp. 234-39. 

14. McRuer, D., I. Ashkenas, and D. Graham. Aircraft 
Dynamics and Automatic Control. Princeton: Princeton 
University Press, 1973. 

15. Lanchester, F.W. Aerodonetics. New York: D. Van 
Nostrand Co., 1909. 

16. "Construction et Manoevure de l'aeaeroplane 
Wright." Monographies d'Aviation No. 5, 1909, Librairie 
des Sciences d'Aeronautiques (F. Louis Vivien, Libraire- 

Aerodynamics, Stability, and Control 


Longitudinal Dynamics of the 
Wright Brothers' Early Flyers 

A Study in Computer Simulation of Flight 



In 1910 when I was five years old, my father took me 
in his 1907 Pope-Hartford out to Simms' Station near 
Dayton, where we lived, to see the Wright brothers 
fly. It was a good day, with several flights. On one of 
them I cried out to my father that the airplane was 
going backward — the little wing was behind the big 
wing instead of in its regular place in front. My father 
smiled and said he thought it was a different kind of 
airplane. I was deeply impressed by the sights and 
sounds. I can still hear the strange clattering sound 
of those old engines. When we got home, I found an 
old orange crate, and put a stick in the ground for a 
control stick, and I had myself an airplane under the 
cherry tree in the backyard. 

Years later, with a record of those early flights, I 
learned that only once had the new model B, with 
its elevator in the rear, ever flown on the same day 
with its predecessors having the elevator in front. 1 It 
was the day of the model B's first flight, May 21, 
1910. It was also the day of Wilbur's last American 
flight as pilot, and I may also have seen that. 

Ten years later as a schoolboy of 15, still possessed 
by airplanes, I became the designer of a machine that 
was being built by a group of schoolmates. Orville 
Wright was a trustee of our school and it seemed 
sensible to talk to him about our design. He received 
us with the same grave courtesy he would have 
accorded any visiting group, and talked to us in grown- 
up terms. We were charmed, and went back to see 

him many times. He loaned us a little fixture he had 
made to shape wing sections of wax, and we made 
wings and tested them in his wind tunnel. He clearly 
enjoyed our visits and was never too busy to see us, 
and we loved him. We talked about all kinds of things. 
He showed us how his desk computer worked in great 
detail and explained the design of the six-cylinder 
Wright engine he was shipping to Canada for his boat. 
He had a wonderful sense of humor and was very 
sharp witted. 

Alone of that group I went back to see him many 
times during later years and we were good friends 
until he died in 1948. I often think about Orville 
Wright and the fun he would have had with a modern 
computer. I can recall very little that we ever talked 
about concerning the longitudinal stability of their 
machines. I do remember his telling that he and 
Wilbur had flown their first (1900) glider backward 
down the hill with only sandbags in it, but I missed 
the entire significance of that remark at the time. 
When he was talking about the superiority of the 
RAF 15 wing section over the USA4 that we had 
chosen, he said, "You will have a lot less trouble in 
controlling it." I didn't understand that either, and I 
didn't ask him to explain. We talked very little in 
later years about flying in the old days, and there are 
hundreds of questions that I now wish I had had the 
sense to ask him while he was living. 

In 1969 I read Charles Gibbs-Smith's The Invention 


of the Aeroplane (Reference 1). It was the best statement 
I had seen of the case for the Wrights as undisputed 
claimants to the honor of being the first to fly in 
controlled powered flight. It had objective evaluations 
of their capabilities and those of rival claimants, and 
I was most favorably impressed. However, it made 
the first mention I had ever seen of the longitudinal 
instability of the early Wright machines and criticized 
the Wrights for persisting with the designs. 

I wrote to Gibbs-Smith at once, expressing my 
appreciation of his work, but taking issue with his 
statement of instability, sending a small flying model 
canard to make my point. He replied modestly saying 
that he was no expert but his expert friends had 
assured him the machines were indeed unstable. 

I couldn't find an accurate technical description of 
the Wright machines to prove my point, and had no 
success in trying to persuade any of my qualified 
friends to undertake a study. It was 1978 before I 
had finally reached the point of deciding I would have 
to do that myself. I knew that having read all the 
books available before 1920 and having taken a few 
undergraduate courses in aeronautics in the 1920s 
didn't qualify me for the job, but I decided that I 
could learn enough to simulate the flight of the old 
machines on the computer in two dimensions only. 

I began with purely linear characteristics, but soon 
found out that it was possible to get into situations 
that demanded real-life non-linearity and aerodynamic- 
data. There were not many laboratory tests of Wright 
airfoils, but I began by using the Wrights' own reports 
of their wind tunnel work, reported in McFarland's 
Papers of Wilbur and Orville Wright (Reference 2). I 
found it surprising that the Wrights had made no 
recorded measurements of center of pressure location 
until 1905, at the time they had finally decided to 
increase the size of the stabilizer on their 1905 
machine, and finally solve their longitudinal control 
problems. There was no correlation between the 
Wrights' wing sections and their center of pressure 
figures, so I used the center of pressure figures of the 
Eiffel 10 (Reference 3) without realizing that in 
combination with the lift coefficients of the Wrights 
they would give an optimistically low value of pitching 
moment coefficient. Even so, there was no doubt the 
machines were unstable, and I made my acknowl- 
edgments to Gibbs-Smith both publicly (Reference 
4) and personally, but I defended the Wrights against 
his charge that they were wrong in continuing the 
unstable designs as long as they did. I argued that 
they had avoided a much worse risk when they 
eliminated the possibility of diving after a loss of 
flying speed. 

Being invited to join a group from the Los Angeles 
Chapter of the AIAA who planned to build a "look 
alike" flying replica of the 1903 Flyer, plus a historical 

replica for wind tunnel testing, I resolved to upgrade 
my obsolete capabilities in aerodynamics and mod- 
ernize my concepts of longitudinal dynamics. Having 
no colleagues nearby with similar interests, I wel- 
comed the prospect of people to interact with. 

The following discussion gives a brief summary of 
the Wrights' careers with major attention to their 
learning and experiences as pilots, and to their canard- 
type Flyers, with major attention to their longitudinal 
control and stability characteristics. Its purpose is 
narrowly confined to the longitudinal dimension only; 
to the plane of symmetry. Fortunately there is neg- 
ligible interaction between the logitudinal control of 
these machines and their lateral or directional controls. 

To accomplish a simulation of their flying behavior 
it has been necessary to establish the aerodynamic 
characteristics of these machines from basic aerody- 
namic theory, since there had been up until recently 
no experimental determination of these characteris- 
tics. It has also been necessary to make a careful 
study of what has been recorded of the flying char- 
acteristics of the machines and their pilots. The flights 
of the 1903 Flyer and the essentially similar 1904 and 
early version 1905 models, which, because of their 
inadequate controls were highly dependent on the 
quality of the piloting, have provided enough infor- 
mation to base a program of pilot simulation. This 
program is able to reproduce some of the historically 
recorded flights with sufficient realism to encourage 
me to believe that they actually reproduce the flight 
characteristics of the Wright brothers and their Flyers. 

I am greatly indebted to Thayer School of Engi- 
neering for countless hours on Dartmouth's superb 
computer system; to E. Eugene Larrabee, who re- 
viewed my early program in detail and made useful 
and constructive criticisms on it; to my West Coast 
colleagues, Frederick E. C. Culick, Henry Jex, and 
Charles McPhail, for useful data, ideas, and com- 
ments; to Tom Crouch, Howard Wolko, and Jay 
Spenser of the NASM staff for assistance in gathering 
data; to Walter P. Maiersperger for useful criticisms 
and comments; and to Candace Dube who processed 
the copy for this report. 

The Wright Glider 

The Wrights learned to fly, developed their control 
concepts, and tested their aerodynamics on their 
gliders. Their fundamental means of longitudinal and 
lateral control, the forward stablizer and the warping 
wing, remained essentially unmodified from the be- 
ginning of their experiments throughout the period 
from 1899 to 1910, but the rudder and its interaction 
with the lateral control continued to evolve during 
the entire glider period, 1900 to 1902, and beyond to 


Frederick J. Hooven 

1904 when the rudder was given its own separate 

They built their 1900 glider (Figure 1) according 
to the aerodynamic tables of Otto Lilienthal so that 
it could be flown as a man-carrying kite tethered in 
the wind at Kitty Hawk, and they planned to teach 
themselves to fly that way. It was a surprise, therefore, 
when they found that their machine would not support 
a man in a wind of 16 miles per hour as they had 
calculated it should. They therefore flew the machine 
as a kite with sandbags on it, and measured the forces 
of lift and drag. They flew the machine backward 
down the hill, finding that it was stable that way, but 
decided they would stick with their forward stabilizer 
to avoid being killed in a dive following a stall, as 
Lilienthal had been. They finally made a few glides, 
finding out if they rigged their machine with a dihedral 
angle to gain stability, they lost a certain degree of 
control and the machine responded to wind gusts. 
Here was born their distrust of stable machines. It 
was not until 1905 that they finally tried a slight 
dihedral angle for improved lateral stability, long after 
Kitty Hawk's wind had been left behind. 

They attributed the lack of lift of the 1900 machine 
to insufficient camber of the wing section, so the 1901 
glider was built with a camber of Viz and a much 
greater wing area. It was almost impossible to control 
and it disappointed them in having not as good a glide 
angle as their 1900 machine. They were greatly 
disappointed with the performance of the glider and 
discouraged to find that Lilienthal's aerodynamic data 
were in error, as were his words about center of 
pressure movement. We are fortunate that E.C. 
Huffaker, representing Octave Chanute, was a visitor 
in the Wrights' camp at Kitty Hawk during the early 
part of the 1901 season in which they were struggling 

with the problems presented by their 1901 glider. 

Huffaker's notes (Reference 5) illuminate the 
Wrights' thinking on the subject of the stability of 
their machine. On July 29, 1901, he noted: "The 
equilibrium is not satisfactory and the Wrights think 
of making radical changes, placing the rudder in the 
rear, or possibly rebuilding the machine. 2 The ex- 
periments have satisfied all of us that as the angle [of 
attack] changes the center of pressure moves forward. 
The curvature of the surfaces while at rest is about 1 
in 20. When carrying a [total] weight of about 240 
pounds the curvature increases to 1 in 12 and is 
constantly changing. To render the curvature more 
constant it is proposed to bring the rear spar forward 
so as to insure less bending of the ribs under pressure" 
(Reference 6). (The ribs were thin wooden strips and 
there was only a single layer of wing covering.) "It 
has been suggested that it would be well to bring the 
rear spar forward so far as to cause the rear of the ribs 
to bend upward under pressure and so lessen the 

That this was in fact done may be seen by looking 
at Huffaker's sketch, Figure 2, of the 1901 wing 

Ir-X-fcL. a • ii . ii ' iia im\ a ' i%\ m ix 

- \ u^U 7 HI It.) IH7 lli]lt.1 



*(.»- w 


jdLj$£. ii£ ios ioj a' j £ji£ io* ii' ioi 

IV ''ill. l o ■ 


3oi ! :U3 

i&5 £ to+t \tffj. a ? /(a/ /L j> ftty ]gg]g£ 

lo ] 1&-\ II I // 

L.i, m 1 7.1 '■ xi 

11 7 /lo' 1)7 1% 

10 10 jj lo' 

"> \ /7J : 1} IU 

tf-XUr'-XH a.h 

A"f f Y r?/U.. 

in nsr hi. at! m 

"i 1J=' I3fi' ijy n.h 

lit x\\- ivriii xin 

10' l«i l 11° ll' Ii' 

2. Huffaker's sketch, taken from his diary, of the 
wing section of the 1901 glider before and after 
camber reduction. Included at top are typical 
Huffaker flight records. 

1. The 1900 glider. (From the Collections of the Library' 
of Congress) 

3. The 1901 glider. (From the Collections of the Library 
of Congress) 

Longitudinal Dynamics 


section before and after the rebuilding operation, in 
which a third spar was added to push down the rear 
of the ribs to flatten them out, and then looking at 
the picture, Figure 3, of the 1901 machine in flight 
after the modification was made. In the photograph 
it is clear that the wing ribs are bent up more in the 
rear in flight than is shown in the sketch. Looking at 
this rib, with its considerable reflex rear edge, it might 
be surmised that the machine had actual positive 
longitudinal stability. That it was indeed so is shown 
by a note made by Huffaker on August 13, after the 
modifications were made. "Fourth experiment, 200 
feet, angle 9° straight and steady, acting automatically, 
i.e. leaving the rudder unchanged." 

The Wrights didn't like Huffaker: he shirked camp 
duties, complained about hardships, and didn't change 
his linen as often as they thought was necessary. But 
we owe him much for his notes, which were much 
better than the Wrights kept. Huffaker's note of the 
"automatic" flight of the glider was not recorded 
elsewhere, nor was his note of July 29, 1901, that 
"Orville Wright was on the machine." There is no 
other record of Orville's flying before 1902. 

The end of the 1901 season was a time of truth for 
Wilbur and Orville Wright. Having found Lilienthal's 
tables to be mistaken, they realized that they would 
have to develop their own aerodynamic information 
if they were to make any further progress in flight, 
and they were deeply discouraged. Finally, they 
decided they were committed to flying. It was the 
turning point in their careers at which they ceased to 
look on their work as a sport and undertook a serious 
commitment to proceed through the development of 
a power-driven machine. 

Their 1902 glider was designed from the results of 
their 1901 wind tunnel experiments, and it was a 
major advance over anything that had ever been built 

before. By far the largest machine ever flown, it had 
a higher aspect ratio and less wing camber than the 
original version of their 1901 glider. The Wrights 
recorded that its wing camber was Vis (Reference 7), 
but it may be seen from photographs to have been 
less than that in flight. This machine saw the final 
development of their ideas of control when its original 
fixed biplane rudder was replaced with a single-plane 
rudder that moved in conjunction with the wing 
warping to counteract the turning effect of the in- 
creased drag on the rising side. Its system of wing 
warping was changed so that only the rear half of the 
outer section of the wing was moved, as may be seen 
in Figure 5, instead of twisting the entire biplane 
structure, as may be seen in Figure 3, of the 1901 
machine in flight. They made more than 1,000 glides 
in this machine and each man had soared in it for 
over a minute. 

Although the Wrights had repudiated the concept 
of shifting weight for glider control, as had been 
practiced by their predecessors, there can be no doubt 
that so long as the operator's weight was greater than 
that of the glider it was utilized for optimizing the 
longitudinal control of their machines. As a result of 
this they were not prepared for the fractious pitching 
behavior of their first powered machines, which had 
crucial differences in wing-rib construction, with the 
first doubled-surface wings, and which were far too 
heavy to be susceptible to trimming by weight- 

The Early Wright Flyers 

The Wrights built five different models of front- 
stabilizer, or "canard" Flyers, the 1903, 1904, 1905, 
1907, and 1909 models. The last was a unique machine 
built for the special purpose of meeting the Signal 

4. The 1902 glider. (From the Collections of the Library 
of Congress) 

5. The 1902 glider banking, showing wing warping. 

(From the Collections of the Library of Congress) 


Frederick J. Hooven 

Corps speed requirements, and had no influence on 
subsequent development, so will not be given further 
attention here. The remaining four are conveniently 
divided into two types aerodynamically: the 1903 and 
1904 machines which differed only in small respects, 
and the 1905 and 1907 models which were substan- 
tially identical to one another. The only important 
difference between these two general types concerns 
the size and operating radius of the stabilizer, which 
was too small for adequate control in the two earlier 
models. The wings of the four were substantially 
identical aerodynamically, differing only in structural 
details and weight. 

The 1905 machine was built similar to the 1903 
and 1904 models, but after it was wrecked on its 
eighth flight, it was rebuilt in the later form. All 
references to the 1905 machine here are made to the 
later form. 

The brief flying life of the 1903 machine has been 
adequately recounted many times. It made only the 
four flights on December 17, 1903, which totaled 
about 95 seconds, before it was wrecked when the 
wind rolled it over. That wasn't much, but it was 95 
seconds longer than anyone had ever flown before. 
The pitching tendency of this machine was very 
troublesome, and at that time it was attributed to the 
overbalancing of the stabilizer. 

The 1903 machine's brief history gives little basis 
for conclusions about its characteristics, but much can 
be learned by studying the flight of the 1904 Flyer, 
which was substantially identical in its specifications, 
except that it had more power, about 16 horsepower 
that gradually grew to 18 at the end of the season as 
the engine became worn in. This machine existed in 
three phases, the first of which was the same as the 
1903, except for about 34 pounds of added weight. 
The second phase followed the rearward movement 
of the center of gravity from 29 percent chord to 32 
percent. This move was made in the expectation that 
it would improve the pitching, or "undulation," but 
it had the reverse effect (Reference 8) and was quickly 
followed by the third phase when a 70 pound weight 
was hung on its stabilizer frame, moving the center 
of gravity forward to 23 percent chord. This reduced 
responsiveness to the elevator and increased the period 
of oscillation in pitch, but it did not cure the problem. 

The 1905 machine was wrecked in a spectacular 
crash on its eighth flight on July 14, 1905, which 
Wilbur described in the following words: ". . . the 
machine began to undulate somewhat and suddenly 
turned down at a considerable angle, breaking front 
skids, front rudder, upper front spar, about a dozen 
ribs, the lower front spar and one upright. O. W. was 
thrown violently out through the broken top surface 
and suffered no injury at all" (Reference 9). On July 
25, Wilbur's diary gives their first recorded measure - 

6. The 1903 Flyer before its flights. (Courtesy 
Smithsonian Institution) 

IA1 11 


7. The famous shot of the first flight, with Orville 
aboard and Wilbur standing at right. (From the 
Collections of the Library of Congress) 

1 1 

/ jHkJ H 

flat }|| 

JU f , 

8. The beginning of the third flight, showing the 
extreme roll. (From the Collections of the Library of 

Longitudinal Dynamics 


ment of center-of-pressure location (Reference 10), 
and on August 24 the machine took to the air for the 
first time with its new and larger stabilizer and front 
structure. Before that date there had been constant 
trouble with what Wilbur called "the tendency to 
undulation that has marked our flights with powered 
machines" (Reference 11). From August 24 onward 
flying developed rapidly. There was no longer the 
tendency to undulate and there were far fewer crashes. 
The year ended with seven flights that totaled nearly 
three hours duration, more than four times the total 
accumulated on their powered machines prior to 
August 24, 1905 (Figure 9). 

This sturdy machine was rebuilt in 1908 with two 
seats, to accommodate two sitting occupants instead 
of the single prone operator, and with a 30-horsepower 
engine. It was taken to Kitty Hawk for flights to 
refresh the Wrights' piloting skills after two and one- 
half years of nonuse, and to accustom themselves to 
the new control system necessitated by the seated 
position. After each man had about 15 minutes on it, 
it was wrecked. Following this, Wilbur went to France 
and Orville back to Dayton, each to prepare for his 
first flight in public before a skeptical audience. 

The 1905 Flyer was the first real airplane: built, 
developed, and flown for more than 39 minutes in a 
single flight before any other powered airplane had 
even left the ground in controlled flight. It is still in 
existence, having been restored in 1948 in a program 
that got under way before Mr. Wright's death. There 
should be a flying replica of it. 

The 1907 machine was aerodynamically similar to 
the 1905 machine but somewhat heavier. At least 

seven of this model were built by the Wrights and it 
is possible that they were not all exactly alike (Ref- 
erence 11). It is shown in Figures 10-13. No authentic 
drawings exist of this machine, but since the first of 
them was begun at the same time as the 1905, it is 
assumed that the wings are identical, and that, except 
for some changes made in the stabilizer actuation, it 
is aerodynamically identical to the 1905 machine. It 
is the machine that amazed the world in 1908 in 
France and in Washington, D.C., when the Wrights 
flew in public for the first time. It is the machine that 
crashed in Washington, D.C., killing Lieutenant 
Selfridge and seriously injuring Orville. In those early 
days, it established a series of records for altitude and 
duration, and, before Orville's accident in Washing- 
ton, the records went back and forth across the Atlantic 
several times. It is the machine on which the first 
formal student training was accomplished, with Tis- 
sandier, De Lambert, and Lucas-Girardville in France, 
Calderera and Savoia in Italy, and Englehard in 
Germany. One of this model is still preserved in the 
Deutsches Museum in Munich, the only one known 
to survive, although a second is said to have been 
recently uncovered in France. 

As I have indicated, and will show later, the position 


9. The 1905 Flyer over Huffman Farm. (From the 
Collections of the Library of Congress) 

10. The 1907 Flyer, Orville aboard; 1908 Signal 
Corps tests. 

1907/1909 WRIGHT FLYER 

11. Drawing of the 1907 Flyer. The draftsman is 
unknown, but authenticity of details indicates that he 
must have had access to the machine. 


Frederick J. Hooven 

of the center of gravity of the canard is critical to its 
stability. It is interesting to note the difference be- 
tween Orville's machine in Washington and Wilbur's 
in France in 1908. The two seats of Orville's machine 
were arranged so that the occupants' legs extended 
for almost their full length forward of the wing's 
leading edge (see Figure 12), while Wilbur's machine 
had the seats so that the knees of the occupants were 
about even with the leading edge. This difference in 
occupant position would make a difference (Figure 
13) of about 3 percent chord in the position of the 
center of gravity of the machine. It casts light on 
Orville's statement to Wilbur in a letter written in 
1908, "I had noticed in my flights with Lt. Lahm 
and Major Squier that there did not seem to be any 
too much surplus pressure on the under side of the 
rudder (stabilizer). It seemed, however, to make the 
flight steadier than when I was on board alone" 
(Reference 13). 

The Wrights have been criticized for persisting as 
long as they did with their unstable canards, a subject 
on which there is much misunderstanding and mis- 
information. The canard is not intrinsically unstable. 
Its stability depends on center of gravity location just 
as in any machine, although the range of stability is 
smaller in the canard. Their avowed reason for the 
use of the canard was to avoid the dive following a 
stall, which they recognized as the cause of the death 
of Otto Lilienthal, their esteemed predecessor, and 
which remains as the cause of many flying accidents 
today. They were well aware that the rear stabilizer 
configuration made it easier to attain stability, and 
they had flown their 1900 glider backward, loaded 
with sandbags, to prove it (Reference 14), but they 
made the deliberate choice of instability. They wanted 
to have the machine go where they directed it to go, 

not where it chose to go, or where the wind might 
take it. They were riders of bicycles and were not 
afraid of instability. 

The canard arrangement (Reference 15) and its 
accompanying longitudinal instability in all probability 
saved the Wrights from serious accident during the 
period, ending in 1905, when they were learning to 
avoid the stall. They repeatedly had trouble with 
stalls in turns, until they realized that turning raised 
the stall speed and that they must put the nose of 
their machine down to hold its speed during a turn. 
They were, therefore, wise to retain that system 
during 1905. They had quite enough to do to build 
more machines and prepare for public flying without 
trying to develop a radically different machine for 
1908, which their critics imply that they should have 
done. The 1907 machine was an almost exact duplicate 
of the 1905 model, except for its greater weight and 
higher power. 

Before they undertook to place their machines on 
general sale in 1910 they developed a new, stable 
machine with the stablizer in the rear, the model B, 
the first flight of which I witnessed. This machine, 
ironically, was involved in a number of fatal accidents 
that resulted from stalling, with the result that the 
model C soon replaced it, with its center of gravity 
farther back and with less stability. 

Simulated Flight 

These simulations are two-dimensional only. They 
are concerned with motion in a single plane, up and 
down, forward, and angularly in pitch angle. It is 
fortuitous that the 1903 flights were all in a straight 
line. This does not mean that because no attempt 
was made to turn the airplane that the lateral control 

: ■ 

12. The 1907 Flyer, Orville aboard, showing forward 
seating position. 

13. The 1907 Flyer, Wilbur aboard, showing seating 

Longitudinal Dynamics 


was unused. On the contrary, as may be seen from 
Figure 8 taken just after the start of the third flight, 
the pilot had more to do than control the plane in 

The aerodynamic specifications of the planes stud- 
ied have been established as shown in detail in 
Appendix 2. 

Over the years in which simulation programs have 
been developed, the general flying characteristics of 
these old machines have shown themselves to be 
encouragingly independent of specific aerodynamic 
assumptions and quantities. This has given me a 
considerable amount of confidence that the simulation 
does represent the way the machines actually flew. 
In this section of the report I will present simulated 
flights under the conditions of December 1903, and 
flights designed to demonstrate the dynamic stability 
of the airplane-pilot combination for the various phases 
of development of the Wright brothers canards. The 
flights demonstrate some of the historic and unique 
flying characteristics such as the "undulation" of the 
1903 and 1904 Flyers and the much lesser, but unique 
continuous small-angle pitching motion of the 1907 
machine, as recorded on film in the famous motion 
picture made at Centocelle in 1909 (Reference 45). 
Finally, these flights demonstrate that the Wright 
canards did in fact remain nearly level after a stall, 
which was why the Wrights chose that form in the 
first place, and why they stuck with it throughout the 
development of their machines. 

Since it is not possible to simulate the flight of an 
unstable machine without some simulated means of 
controlling it, I will discuss the problem of simulating 
the airplane pilot. 

The Wright Brothers as Pilots 

Wilbur and Orville Wright not only invented the 
airplane, they also invented the airplane pilot and the 
very concept of airplane control. Until their widely 
disseminated publication of 1901 (Reference 16), the 
airplane was conceived as a craft that, like an airship, 
would be steered right and left with a rudder, and up 
and down with a horizontal rudder, by anyone who 
happened to be aboard, without previously acquired 
skills. It was what Charles Gibbs-Smith called the 
"chauffeur mentality," and it was the universal state 
of mind before the Wrights' published work, and the 
general view of would-be aviators. Their concept of 
controllability was not understood or accepted until 
their public flights in 1908. 

Builders and riders of bicycles, the Wrights con- 
ceived the airplane from the very first as a craft that, 
like the bicycle, depended on its rider to maintain its 
equilibrium. It was therefore perfectly natural for 
hem to assume that in order to achieve powered 

flight, it would first be necessary not only to develop 
the airplane, but also to develop its controls and the 
skills necessary to operate them. From the beginning 
they assumed that the means of control of their 
machine would be by movement of its surfaces. 

I make no distinction between one Wright brother 
and the other, not only as pilots but in any of their 
contributions to flight. The basis exists for comparing 
their abilities to those of pilots of a later generation. 
I have known three of the men who received their 
flight instruction from Orville well enough to have 
heard their opinions of him as a pilot. They all spoke 
of him as a superb pilot with lightning-quick reflexes 
and sure responses. While no such basis for comparison 
exists where Wilbur is concerned, there can be no 
doubt he was Orville's equal as a pilot, as may be 
seen from the records of their flights as they have 
been recorded by Renstrom (Reference 17). 

The Wrights acquired their flying skills on their 
gliders, and they needed all their skills to control the 

1903 machine on the four flights it made on December 
17 at Kitty Hawk. They flew into a gusting wind that 
averaged about 20 miles per hour, two-thirds of the 
flying speed of their machine. They attributed the 
difficulties with excessive pitching to the fact that the 
elevator was overbalanced, but when they built the 

1904 machine with the elevator pivot point farther 
forward, the pitching persisted. 

The records revealed that their skills were not quite 
up to the task of controlling the 1904 machine or its 
short-lived 1905 successor of similar proportions. In 
these machines there was a total of 90 recorded actual 
flights, out of which there were 19 instances of landings 
with damage to the machine. As nearly as can be 
judged from the sparse records, 8 of these were the 
result of pitch-control inadequacies, 6 from lateral 
control, and 5 of which no clue survives. 

Of 37 recorded flights with the remodeled 1905 
machine, there were only two that resulted in recorded 
damage, the third and sixth after the rebuilding. In 
the last 31 flights there was no further damage, and 
23 of these flights racked up more than four times as 
much flying time as in all the flights of the predecessor 
models. It was without doubt a help that the 1905 
machine had 20 horsepower available and could there- 
fore fly at a somewhat higher speed, 50 feet/second 
(34 miles per hour) instead of 44 feet/second (30 miles 
per hour). Its performance was sufficient that the 
Wrights decided it could be put on sale. 

No discussion of the Wrights as pilots is complete 
without reference to their system of flight controls. 
Up to the end of 1905 all their flying in glider and 
Flyer had been made in the prone position, using the 
hip girdle to actuate the combined wing warping and 
rudder control, and a left-hand lever with fore-and- 
aft movement to actuate the elevator. Later in 1905 

Frederick J. Hooven 

they separated the rudder from the wing warping and 
added a right-hand lever with fore-and-aft movement 
to actuate it. 

When they started to fly in a sitting position in 
1908 they rebuilt the trusty old 1905 Flyer with a 
new 30-horsepower engine and two seats on the wing. 
The controls were new and different, since the hip 
girdle was no longer used. Instead, a second direction 
of motion was added to the right-hand lever, which 
was moved right and left to control the wing warping 
in addition to moving fore-and-aft to control the 

Each brother got about 15 minutes on that machine, 
including one trip each with a passenger, the first 
ever carried, before it was wrecked. Therefore, it was 
with but 15 minutes of flying experience with the 
new control system after two and one-half years of 
not flying at all that Wilbur undertook their first flight 
in public, in France before a critical and skeptical 
audience, and Orville shortly after made their first 
public flight in America. After having made three 
short flights in France and wrecking the machine on 
the third landing, Wilbur wrote to Orville, "I haven't 
yet learned to operate the handles without blunders" 
(Reference 18). 

Wilbur and Orville Wright were the first ever with 
"the right stuff." They needed every bit of it that 
summer of 1908. 

The Simulated Pilot; Simulating the Wright 

It is meaningless to speak of the dynamic stability of 
a statically unstable system, and the longitudinal 
dynamic stability of the Wright canards is evaluated 
in combination with a simulated pilot, the responses 
of which are intended to reproduce those of the 
Wrights as nearly as can be deduced from the historical 

I began by assuming that any algorithm that repro- 
duced a plausible performance would be satisfactory 
as a pilot, and for the sake of convenience I chose to 
use airspeed as the target variable, with a response to 
angular velocity in addition. This algorithm was: 

AB = Q2 (V-Ql) - Q4w 

in which Ql is the target airspeed, Q2 was adjustable 
from .015 to .03 and Q4 from 30 to 60. 

There was a reaction-time delay that could be 
varied from to .08 seconds, of which a typical value 

14. Equilibrium values of stabilizer, angle of incidence vs. angle of attack, 1903 Flyer. 

Longitudinal Dynamics 


15. Equilibrium values of stabilizer, angle of incidence vs. angle of attack, 1905 Flyer. 

was .04. 

This controlled the aircraft in a realistic manner, 
but it was criticized by Eugene Larrabee (Reference 
19), who noted that real pilots do not in fact use 
airspeed as their basic response stimulus, but rather 
seek to maintain the pitch attitude of their machine 
relative to the horizon. Acknowledging the fairness 
of this comment, but also recognizing that the pilot, 
when faced with irregular air currents or irregular 
engine output, would be in trouble if he did not also 
pay attention to his angle of attack, or airspeed, I 
devised a new algorithm that used the pitch angle of 
the machine as its target value: 

AB = Q2(Q1-Ap-Ai) 

But whenever the angle of attack deviates from Q\ 
by more than 1 degree the pilot responds to angle of 
attack as well as to pitch angle: 

AB = Q2(Ql-Aa/2-(Ap + Ai)/2) 

The simulated pilot senses, in addition, angular ve- 
locity and angular acceleration only after the angular 
velocity exceeds H — .02 rad/sec and then it adds the 
following correction to that already made: 


Q4w - Q6P 

Experimentation with sensitivity factors led to the 
following values: 

Q2 = 1.8, Q4 = 2.0, Q6 = .002 

to give flying characteristics like those of the historical 
record, when flown with a reaction-time delay of .06 
seconds. The reaction time proved to be easy to 
select. In fact, .04 seconds provided flight control that 
was clearly more powerful than that of the human 
pilot, while .08 seconds gave a pilot that had no 
trouble with the 1905-1907 Flyers, but could not 
control the 1903-1904 Flyers adequately. 

Orville Wright related an occasion, early in his 
flying career, in which he became so engrossed in 
lateral control problems that he forgot altogether about 
his longitudinal control and got into an extreme stall 
position (Reference 46). It is easy to see, however, 
how even an experienced pilot, with a machine whose 
longitudinal control requires the constant reflexes as 
sharp as those of a baseball hitter, can forget himself 
long enough to lose control of a machine whose lateral 
or directional control might be demanding attention, 
as so often happened in the days of the 1903-1904 


Frederick J. Hooven 

The control algorithm resolves into the following, 
neglecting the offsets of W and Aa: 

AB = Q2(Q1-Ap)-Q4 
AB = 1.8(Q1-Ap) -2.0 





d 2 Ap 

- .56 





1904 ph 2 


1904 ph 3 







d 2 Ap 





Concluding Remarks; The Simulated Flights 

In the following, Figures 16-21 represent the flight 
of the 1903 Flyer under conditions typical of Kitty 
Hawk on December 17, 1903. Figures 22-27 represent 
the flight of the 1903, the 1904 phase 2 and the 1904 
phase 3 Flyers, and the 1905 Flyer following pertur- 
bation by a one-second elevator pulse followed by 
normal piloted control as described. Figure 28 shows 
the flight of the 1905 Flyer in a stalled descent, and 
Figure 29 shows the 1905 Flyer in its stable regime 
after certain modifications. 



ta His EHs G£* Q£*. . ras "ffl^ EEs.. 



100 200 300 


~.tSs CBs ....f3s....SSs <3s ^ m 

16. Simulated flight under typical Kitty Hawk 
conditions. Rough landing after 15 seconds, 1903 

The Kitty Hawk Flights (Figures 16-21 ) 

On December 17, 1903, the Wrights made four flights. 
The first three averaged 10 ft/sec over the ground, 
and lasted approximately 12 seconds, while the last 
averaged 14.5 ft/sec over the ground, lasting for 59 
seconds. All were made against winds that averaged 
about 34 ft/sec and which were gusting. "Undulation" 
was a problem, accentuated by the fact that the Flyer's 
elevator was overbalanced and tended to go to ex- 
tremes of travel. 

The simulated flights were all made at an airspeed 
of 44 ft/sec against winds that gusted up to 8 ft/sec 
for a duration of up to 3 seconds, both figures being 
random numbers. Assumed engine power was initially 

is ras ras ras ras ras... ..ras. 

.CBS G3S. 

gas ras ras ras ras ras ras ras ras ras. 

...ras ras ras.....ffls.....ras.....e2s. ....ras.. 

™yce s 

17. Simulated flight under typical Kitty Hawk 
conditions. Rough landing after 28 seconds, 1903 

15, dropping off during the first 10 seconds to 13.8, 
the steady-state power required for level flight at 44 
ft/sec. Altitude is plotted against horizontal distance, 
the distance being air distance except for Figure 21 
where it is over-the-ground distance, assuming a mean 
wind speed of 30 ft/sec. Altitude is plotted at double 
the scale of horizontal distance. The diagrammatic 
graphic Flyers are shown once for each second of 


Feet Altitude 

is ras ....ras ras.. ras ras ^^ 


ras ...ras. ras.. .ras 133s ras ras 

-.■■ca-i . . 

18. Simulated flight under typical Kitty Hawk 
conditions. Landing after 16 seconds, 1903 Flyer. 

flight time, and their pitch angle and angle of their 
elevator planes are represented as they actually were. 
The dots indicate the position of the Flyer at the .1 
second points. 

All flights were made under identical conditions 
except for the random gusts. It will be noted that of 

cost ras ras ras ras ras ras ras ras.. 

.Has ras ras ras ras tS3-...ras--<a&f.. < 

19. Simulated flight under typical Kitty Hawk 
conditions. Crashing after 17 seconds, 1903 Flyer. 

Longitudinal Dynamics 


20. Simulated flight under typical Kitty Hawk 
conditions. Landing after 16 seconds, 1903 Flyer. 

the first five flights, four encountered "undulation" 
before they ended. Of this five, two had normal 
landings, two had rough landings, and one ended in 
a crash. A crash is a vertical velocity exceeding 12 ft/ 
sec, a rough landing exceeds 8 ft/sec. Figure 21 shows 
eight flights, of which four were normal landings, two 
were rough landings, and two were crashes. 

Wind gusts were simulated by arbitrary variations 
in airspeed, and each gust was followed by one of 
opposite sign and similar duration so that mean gusting 
was zero. Gusts strike wings and elevators in the 
proper sequence, causing the pitch angle to be per- 
turbed, and initiating undulation. 



TN-rn umn nF ia rr/iFr utt h »»]nnM raiyr<i up Tn a ft/sfc 

_! ' ' I l_] 1 ' ' i I I I 1 I 1 I I l I_I I L_I — 1 — L-J — I — 1_I — 1 — !_ 

-Feet, Altitude 

ling rate, degreea/sf 


qjs, t3s ijfcs B3fc} djni ojsi las 

JJs, VQ3 Ote UN U3£3 Cff3 EQnj ICpsj CEL} (3s 

PJNi Djni ICTJni Qis Q£i 13^ t33 »3 

22. Simulated flight of the 1903 Flyer after 107sec 
perturbation. C g = 29% chord, 1-second elevator pulse, 

In these tests altitude is plotted against horizontal 
distance, with altitude being plotted at double the 
scale of horizontal distance. The dotted line, however, 
indicates the angular velocity of pitching rather than 
the machine's altitude. As in the above, the graphic 
Flyer appears once for each second of flight. 

The 1904 phase 1 Flyer behaved so much like the 
1903 Flyer that separate tests were not deemed 
necessary. The phase 2 test shows that by moving 


hing Velocity 



OiNa "Sts 












21. "Kitty Hawk" flights, 1903 Flyer. 

23. Simulated flight of the 1904 Flyer, phase 2 after 
107sec perturbation. C g at 32% chord, 1-second 
elevator pulse = .79.° 

The Stability Flights (Figures 22-28) 

Each flight begins with a fixed elevator position for 
its first second. The elevator position was proportioned 
such that a 107sec pitching rate was established, 
following which the normal pilot took over control. 
In the instance of the 1904 phase 3 Flyer, with its 70 
pound ballast weight on the elevator frame, a 107sec 
pitch rate caused amplitudes of such magnitude in 
pitch that the linear range of elevator lift was ex- 
ceeded, so a second test was conducted in which the 
elevator pulse rather than the pulse rate was equated 
with the tests on the other machines. 

- Pitching t.te, 



cs •a* 









" <*.. . 

24. Simulated flight of the 1904 Flyer, phase 3 after 
107sec perturbation. C g = 23% chord. (70 lb. ballast 
on front elevator frame), 1-second elevator pulse = 


Frederick J. Hooven 

the center of gravity from 29 percent to 32 percent 
chord the Wrights put this machine beyond the limits 
of controllability. 

The 1904 phase 3 test shows the effect of the great 
increase in moment of inertia brought about by the 
addition of the 70 pound ballast weight on the elevator 
frame, an increase from 295 si ft 2 to 445 si ft 2 . The 
107sec pitching rate puts the elevator beyond its range 
of linear lift, and the unsymmetrical pitching-rate 
curve shows that the elevator is approaching the stall 
point during the oscillation. The result of this is a 
radical reduction in damping with a catastrophic in- 
crease in amplitude of pitch. 

25. Simulated flight of the 1904 Flyer after 
perturbation by elevator pulse of .78° for one second. 
C g = 23% chord, phase 3 with 70 lb ballast weight. 

Testing the phase 3 Flyer, with an elevator pulse 
equal to that impressed on the 1903 and 1904 phase 
2 Flyers, shows an oscillation within the limits of 
linear lift for the elevator. This is the state in which 
the 1904 Flyer spent most of its flying life. 

Tests of the 1905 Flyer with a 107sec initial 
perturbation show quick damping and a rapid reduc- 

1 " Pltc 

lng Velocity, 

de 8 „,e. 

D - 

j- .1.1 

ude, feet 

Cjfcal E35S 









QIS3 CPv? irmva 






a*> g fcg a^ eQja cffiy as o^ ^ jgj. 

26. Simulated flight of the 1905 Flyer after 107sec 
perturbation. C g = 18% chord. 

Longitudinal Dynamics 

tion in amplitude of pitch, with a small residual 
pitching rate that was characteristic of the 1905 and 
1907 machines (Reference 45). 

Figures 27 and 28 show the effect of an increase 
of .02 second in the pilot's reaction time delay with 
the 1903 and 1905 Flyers. A .06 second time delay is 
in the same range as those required of a baseball 
hitter, and an increase in that time from .06 to .08 
seconds results in a loss of control of the 1903 Flyer, 
as shown by Figure 27. A similar increase has almost 
no detectable effect on the 1905 machine. 


cftj tSs t2s CBs Eflj eSs 


27. Simulated flight of the 1903 Flyer after 107sec 
perturbation, with pilot reaction-time delay of .08 
seconds. C„ = 29% chord. 

,ts* Ob* 

Cffia, Os as, ca^ c^ ^^ ^^ 

CBS=J CBfcj 

cbs* eas3 Cass la^ 


a» q^ q^ .^ ^ O^ c^, ^ 

28. Simulated flight of the 1905 Flyer after 107sec 
perturbation, with pilot reaction-time delay of .08 

The Stalled Descent of the 1905 Flyer (Figure 29) 




, HV» 

29. Simulated flight of the 1905 Flyer illustrating stall 


The entire rationale of the Wrights' use of the forward 
elevator is shown by this flight, which required that 
the pilot be directed to watch pitch angle only. The 
machine does descend without diving, its angle of 
attack at landing being 48°. Even so, from 20 feet 
altitude, it strikes the ground at a vertical velocity of 
13 ft/sec, which constitutes a crash. 

The Flight of the Modified Flyer 1905 (Figure 30) 

For the 1905 Flyer, being my sentimental favorite of 
all airplanes, I illustrate the effects of a modification 
similar to those being brought about in certain so- 
called replicas of Wright Flyers that have taken the 
air of late, that of a great increase in power over the 
prototype. The 1905 Flyer is not blessed with low 
drag; its drag is the equivalent of a barn door size 5 
x 9 feet, and, while its historic 20 horsepower would 
propel it at 34 miles per hour, no less than 250 
horsepower would be required to propel it as fast as 
80 miles per hour, at which speed its drag is almost 
equal to its weight. Endowed with thrust of this 
magnitude many things are possible, one of which is 
shown in Figure 30. 



30. Simulated flight of "modified" 1905 Flyer. 

Frederick J. Hooven was Professor of Engineering at the 
Thayer School of Engineering of Dartmouth College. He 
studied Aeronautical and Mechanical engineering at MIT, 
receiving his B.S. in 1927. He then worked for the General 
Motors Research Laboratory and the Army Air Corps. For 
much of his career, he worked as an independent consultant 
and inventor and collaborated in the development of the 
first successful heart-lung machine. In 1956 he joined the 
Ford Motor Company as an executive engineer for advanced 
automotive products. 

Professor Hooven was a member of the National Academy 
of Engineering, American Association for the Advancement 
of Science, American Institute of Aeronautics and Astro- 
nautics, Dartmouth Engineering Society, Dayton Engineers 
Club, Federation of American Scientists, and Sigma Xi. He 
held thirty-eight U.S. patents and published numerous 
technical articles. 

Frederick J. Hooven died before this work was published. 


1. The term "elevator" is used interchangeably with 
"horizontal rudder," "horizontal stabilizer," "stabilizer" 
and "canard." The latter is incorrectly applied to the 
elevator, being the term that is applied to the machine 
whose elevator is in front of the wing. 

2. Whenever Huffaker refers to "rudder" he means 
horizontal rudder, or stabilizer, since the 1901 machine 
had no vertical rudder. 


1. Gibbs-Smith, Charles H. The Invention of the Aeroplane. 
New York: Taplinger, 1965. 

2. McFarland, Marvin W. The Papers of Wilbur and Orville 
Wright. New York: McGraw-Hill, 1953, p. 547. 

3. "Aerodynamic Characteristics of Airfoils," NACA 
Technical Report No. 93, 1919, p. 318. 

4. Hooven, Frederick J. "The Wright Brothers Flight- 
Control System," Scientific American 239:5, November 
1978, p. 167. 

5. Chanute-Huffaker Diary, July-August 1901, pp. 154, 
161, 167. Papers of Wilbur and Orville Wright. Library of 

6. Wright, Wilbur, "Some Aeronautical Experiments." 
Journal Western Society of Engineers, December 1901. In 
this paper Wilbur describes experiments conducted with 
the upper wing of the 1901 glider in the wind at Kitty 
Hawk, in which it was proved that the center of pressure 
of the wing moves forward as the angle of attack 
increases, in contradiction to Lilienthal's data. 

7. See McFarland, p. 253, letter from Wilbur to George 
A. Spratt, September 18, 1902. 

8. The rearward movement of the center of gravity of the 
1904 Flyer in the expectation of improving the pitching 
tendency has been called a blunder by some, including 
myself, but I have changed my opinion. The Wrights 
were not expecting to improve the static stability of the 
machine, and there is no doubt they were entirely 
familiar with static stability inasmuch as they were 
acquainted with the work of Penaud and they had made 
and flown many models. They knew that moving the 
center of gravity to the rear would increase the machine's 
response to stabilizer movement, and it was entirely 
reasonable to suppose that this move would help the 
pitching difficulty. 

9. Diary F of Wilbur Wright, 1905, pp. 6-7. Papers of 
Wilbur and Orville Wright. Library of Congress Archives 

10. Notebook H of Wilbur Wright, p. 16. Papers of 
Wilbur and Orville Wright. Library of Congress Archives 

11. See McFarland, p. 445, letter from Wilbur to Octave 
Chanute, July 17, 1904. 

12. The drawing, Figure 11, is not authenticated, but it 
checks in all known details and dimensions with its 
subject, and the original drawing must have been made 
by someone with access to the machine itself. 

13. See McFarland, p. 938, in a letter from Orville to 
Wilbur, November 14, 1908. 

14. Personal communication from Orville Wright (see p. 
3). This is the only personal communication from Mr. 
Wright I can cite in relevance to the current subject of 
longitudinal dynamics. 


Frederick J. Hooven 

15. It is a fact that any machine, whether a canard or not, 
will descend more or less flat in a stall if its center of 
gravity is far enough back, but the forward placement of 
the stabilizer makes this fact intuitively more apparent. 

16. See Wright, Wilbur, "Some Aeronautical 
Experiments." This paper was widely published in 
Europe and is credited with reviving interest in flying, 
especially in France, where the first gliders and powered 
machines clearly derived from the Wrights. 

17. Renstrom, Arthur G. "Wilbur and Orville Wright, a 
Chronology. " Washington, D.C.: Library of Congress, 

18. See McFarland, p. 912, letter from Wilbur to Orville, 
August 15, 1908. • 

19. Larrabee, E. Eugene, private communication with 
the author. 

20. Munk, Max M. "General Biplane Theory." NACA 
Technical Report No. 151, 1922. 

21. Glauert, H. Elements of Airfoil and Airscrew Theory, 
MacMillen 1943, p. 125. 

22. See Glauert, p. 87. 

23. Drawings of 1903 Wright brothers Flyer. National Air 
and Space Museum, Smithsonian Institution, Washington, 

24. Drawings of restored 1905 Wright brothers Flyer, 
U.S. Air Force, Wright-Patterson Air Force Base, Dayton, 

25. Abbott, Ira H., and von Doenhoff, Albert E. Theory 
of Wing Sections. New York: Dover Publications, 1959, pp. 

26. Theodorsen, Theodore. "On the Theory of Wing 
Sections, with Particular Reference to Lift Distribution." 
NACA Technical Report No. 383, 1931. 

27. Knight, M., and Wentzinger, C.J. "Wind Tunnel 
Tests on a Series of Wing Models through a Large Angle 
of Attack Range." NACA Technical Report No. 317, 

28. Bettis, W.H., and Culick, F.E.C. "Wind Tunnel 
Tests on a 14 Scale Model of the Wright Brothers 1903 
Flyer Airplane." Guggenheim Aeronautical Laboratory, 
California Institute of Technology Report No. 1034, 
Pasadena, Calif., 1982, p. 20. 

29. McPhail, CD. "Textbook Analysis of the 
Longitudinal Characteristics of the 1903 Wright Flyer." 
Private communication, 1981, p. V-l. 

30. Wood, K.D. Technical Aerodynamics, New York: 
McGraw-Hill, 1935. 

31. "Aerodynamic Characteristics of Airfoils, V." NACA 
Technical Report No. 286, 1928, REF 649. 

32. See Abbott, et al., p. 111. 

33. Millikan, Clark B. Aerodynamics of the Airplane. 
London: Wiley, 1941, p. 47." 

34. Silverstein, A., Katzoff S., and Bullivant Kenneth W. 
"Downwash and Wake Behind Plain and Flapped 
Airfoils." NACA Technical Report No. 651, 1938, Figure 

35. Notebook K of Wilbur and Orville Wright, 1903. 

Papers of Wilbur and Orville Wright. Library of Congress 
Archives Folder. 

36. Orville Wright to Charles Taylor, November 23, 
1903. The Papers of Wilbur and Orville Wright. Library 
of Congress Archives Folder. 

37. See McFarland, Appendix pp. 1212, 1213. 

38. Chenowith, Opie. "Power Plants built by the Wright 
Brothers." SAE Quarterly Transactions 5, January 1951, pp. 

39. Noyes, Richard W. "Pressure Distribution Tests on a 
Series of Clark Y Biplane Cellules with Special Reference 
to Stability." NACA Technical Report No. 417, 1932. 

40. Munk, Max M. "Air Forces on a Systematic Series of 
Biplane and Triplane Cellule Models. NACA Technical 
Report No. 256, 1926. 

41. Crowley, J.W., Jr. "Pressure Distribution over a 
Wing and Tail Rib of a VE-7 and of a TS Airplane in 
Flight." NACA Technical Report No. 257, 1927. 

42. The first of the 1907 Flyers was built along with the 
1904 and 1905 machines in the spring of 1904 (see 
McFarland, p. 442, note 1), which makes it unlikely that 
it had a unique wing section. The drawing in Figure 11 is 
of unknown origin and unauthenticated, but it checks 
what is known of that machine, and it shows a rounded 
leading edge similar to those on other Wright machines. 
It has therefore been concluded that, at least in this 
respect, the Eiffel 10 is not a representative Wright wing 

43. The Bettis and Culick report shows that the 
combination of wing and stabilizer has zero lift at —3.4° 
angle of attack with the stabilizer incidence being 0. It 
has been assumed that both sections exhibit under-wing 
burbles, and that their true angles of zero lift will bear 
the same proportion as their computed values of ao and 
ao h , which are —4.84 and —2.8 respectively. Since the 
sum of both lifts is at —3.4° it can be said that 


= - 

Kh and 


3.4 - 

aoh') = .063(3.4 - ao' 

3.4 ao h ' = 


io' - 33.422 




at h • 

- ao, 

- ao 

,' at h 

— ao t 

- ao 


.3 + 2.8 


1.0828 + 4.84 


,' = .2667 + 


ao' = 36.822 - 

- 9.83ao' 

then ao = —3.58 and ao h = —2.14 

44. Wilbur to Orville, August 9, 1908. Papers of Orville 
and Wilbur Wright. Library of Congress, Archives Folder. 

45. This effect is seen in the first moving picture ever 
taken from a moving airplane, a Bioscope cinema of 
which a copy resides in the National Air and Space 
Museum, Washington, D.C. 

46. Diary B of Orville Wright, September 20, 1902, in 
which he describes forgetting to operate the elevator in 
his concentration on a problem with the lateral control. 
The Papers of Wilbur and Orville Wright. Library of 
Congress, Archives Folder. 

Longitudinal Dynamics 


Table 1 

Weight per Unit Wing Area, 


1.79 lb 

Dimensions of 1903 and 1904 Flyers 

1907 1 
1907 2 


2.04 lb 

Wing Span 

40.33 ft 

2.32 lb 

Wing Chord 

6.5 ft 



8.5 ft 

Wing Gap 

6.0 ft 

Wing Area 

510 sq ft 

Wing Aspect Ratio (2b 2 /S) 


Table 3 

Elevator Span 

12.0 ft 

2.54 ft 

Thin-Wing Functions 

Elevator Chord 

Elevator Gap 

2.17 ft 






Elevator Area 

52.75 sq ft 






Elevator Aspect Ratio 







Elevator Moment Arm from Leading 







7.32 ft 






Rudder Height 

7.0 ft 






Rudder Length (chord) 

1.5 ft 






Rudder Area 

10.26 sq ft 






Rudder Moment Arm from Leading 







11.5 ft 






Cg Position from Leading Edge 






1903, 1904 ph.l 

1.89 ft, 29% 






1904 ph. 2 

2.21 ft, 34% 






1904 ph. 3 

1.42 ft, 21.8% 






Flving Weight 







750 1b 






1904, ph. 1,2 

784 lb 






1904, ph. 3 with 70 lb wt 

854 1b 






Weight per Unit Wing Area 







1.47 lb/sq ft 






1904, ph. 1,2 

1.54 lb/sq ft 






1904, ph. 3 

1.67 lb/sq 






Propeller Diameter 

8.5 ft 































Table 2 






Dimensions of 1905 and 1907 Flyers 











Wing Span 

40.33 ft 

\J • -J^J \J\J 


Wing Chord 

6.5 ft 






Wing Gap 

6.0 ft 






Wing Area 

510 sq ft 






Wing Aspect Ratio (2b 2 /S) 







Eleavator Span 

15.63 ft 






Elevator Chord 

3.0 ft 






Elevator Gap 

3.0 ft 






Elevator Area 

83 sq ft 






Elevator Aspect Ratio 





Elevator Moment Arm from Leading 







11.7 ft 






Rudder Height 

7.0 ft 






Rudder Length (chord) 

2.25 ft 






Rudder Area 

15.5 sq ft 






Rudder Moment Arm from Leading 







11.18 ft 






Cg Position from Leading 






Edge 1905 

1.22 ft, 18.8% 






1907 1 man (1) 

1.0 ft, 15.32% 






1970 1 man (2) 

1.1 ft, 17% 






1907 2 men (1) 

.87 ft, 13.34% 






1907 2 men (2) 

1.07 ft, 16.39% 






Flying Weight 1905 

914 1b 






1907 1 man 

1041 lb 






1907 2 men 

1186 lb 






Frederick J. Hooven 











































3. 1494 

























































































1546 . 



































































































Table 4 

Pankhurst's constants and upper and lower ordinates of the 
1903 wing listed against chordal station. 
X A B L U 

































































Table 5 

Coefficients of lift, drag, pitching moment for the wing and 
lift for the uncambered stabilizer section, listed for angles 
of attack from - 4 to + 30. 
a CI Cd Cm Kl* 




























* Stabilizer lift coefficient is referred to the wing area for 
the 1903 Flyer. 

Table 6 

Mean line, trailing section of 1903 rib, deflection with 
maximum of 1% at trailing edge. The chord line is drawn 
for the deflected rib, so the undeflected coordinates are 
shown below the chord line. 

Undeflected Deflected Deflection 

X Ordinates Ordinates (vXlO 5 ) 




























































































Longitudinal Dynamics 



A = Effective monoplane aspect ratio 

A = Pankhurst's integration constant for Ao 

Aa.aa = Angle of attack, radians and degrees 

Ab,ab = Absolute angle of attack (Aa — Az) radians and 

Af,af = Flight path slope angle, radians and degrees 
Ah, ah = Incidence angle of stabilizer = Incidence angle of wing 
Ao,ao = Angle of zero lift 

Ao',ao' = Effective value of ao with under-wing burbles 
Ap,ap = Airplane pitch angle 
At, at = Theodorsen's "ideal" angle of attack 
B = Pankhurst's integration constant for Cq 
B = Stabilizer-to-wing angle, degrees + 10 = Ah — Aa 

+ 10 
B = Munk's ratio of 2-D to 3-D biplane lift 
Be = Effective stabilizer angle corrected for pitching rate 
B3 = Automatic-pilot reaction-time delay factor 
B7 = Automatic pilot offset for pitching rate correction 

Cd = Drag coefficient for entire airplane except stabilizer 
Cg = Center of gravity longitudinal coordinate, percent 

CI,L(I) = Lift coefficient for wing 
Cm = Pitching-moment coefficient 
Cmac = Pitching-moment coefficient about aerodynamic 

C = Pitching-moment coefficient about Cg 
Cq = Pitching-moment coefficient about quarter-chord 

Dg = Drag, pounds for complete airplane except 

DL = Distance from leading edge to stabilizer Vi chord, 

percent chord 
Ed = !/2da*S, aerodynamic factor 
E(I) = Automatic-pilot time-delay count 
Cv = Random-wind maximum velocity for gusts ft/sec 
H = Engine-shaft horsepower, assuming propulsion 

efficiency of 72.7% 
Im = Airplane longitudinal moment of inertia, slugs feet 2 
1,12 = Angle of attack +10 and fractional degrees for 

data matrix 
J,J2 = Stabilizer angle +10 and fractional degrees for 

data matrix 
K1,K(I) = Stabilizer lift coefficient for uncambered 

Kh = Stabilizer lift coefficient corrected for camber 
Lz = Total airplane lift for stabilizer and wing, pounds 
L = Lower-surface ordinate for airfoil section, chord/100 
N = Effective camber reduction factor for under-wing 

N = Elapsed time count of iteration period in computer 

P = Pitching moment, pounds feet 
Pd = Dynamic pressure, SV 2 da/2 
Q = Input variable for program control codes 
Ql = Target value of angle (ai + ap + af )/2 for 

automatic pilot, degrees 
Q2 = Sensitivity index for pilot for angles of pitch and 

Q4 = Sensitivity index for pilot for pitching motion 
R = Ratio of wing area to stabilizer area 
R = Theodorsen's value of (x(l — x)) 1 '.' 
S = Wing surface area, square feet 
f = Elapsed time of flight, seconds 

Th = Thrust pounds 

U,U1 = Stabilizer angle of attack, degrees +10 and 

fractional degrees 
U = Upper-surface ordinate for airfoil section percent 

V,Vh = Airspeed, feet per second, at wing and at 

Vv = Vertical component of velocity of airplane 
Wt = Weight of airplane, pounds 
W,w = Angular velocity of pitching motion, degrees and 

Y = Thin-wing airfoil maximum ordinate, percent chord 
Yh = Stabilizer section maximum ordinate, percent chord 
Z = Altitude of flight, feet 
b = Wing span, feet 
ch = Wing chord, feet 
cp = Center of pressure, percent chord 
dV = Airspeed increment, ft/sec/sec 
da = Air density, slugs/cu ft 
e,u = Variables used in computation of differential lift 

g = Wing gap, feet 

p = Power of si in lift-coefficient equation 
p = Differential coefficient of profile lift 
q = Differential coefficient of induced lift 
r = Radius of airfoil section leading edge, percent chord 
SO, so = Lift-curve slope, of 2-D biplane per radian and 

per degree 
SI, si = Lift-curve slope, of 3-D biplane per radian and 

per degree 
t = Iteration time for computer program, seconds 
x = Distance along chord line, percent chord 
xp = Value of x as transformed for profile lift 
xq = Value of x as transformed for induced lift 
xe = Value of x associated with given value of e 
y = Ordinate of mean-line thin-wing airfoil section, 

percent chord 
yh = Ordinate for stabilizer section, percent chord 
ye = Effective value of y as reduced by under-wing 

zl = Vertical coordinate of cener of lift, percent chord 
z2 = Vertical coordinate of center of thrust, percent 

z3 = Vertical coordinate of center of drag, percent chord 
as subscript p applies to profile lift 
as subscript i applies to induced lift 
as subscript h applies to stabilizer. 

Frederick J. Hooven 

Appendix 1 

The Wright Wing Section 

The five Wright canards all had wing sections of the 
same general shape, thin, highly cambered, tapering 
from front to rear, yet all differing in small ways. The 
1904 Flyer, destroyed without drawings ever having 
been made, was reported by the Wrights to have a 
camber of Vzs instead of the approximately V2.0 of the 
other machines. Replica ribs of the 1903 and 1905 
machines were built in the course of this investigation, 
as shown in Figure 31. The 1905 rib is also probably 
typical of the 1904 and 1907 models since all three 
were begun at the same time in the spring of 1904. 
These two airfoils are aerodynamically alike. 

The Eiffel 10 section is labeled "Wright," and it 
resembles the 1903 and 1905 sections, except for its 
leading edge which is a vertical line with sharp corners 
at the intersections of the upper and lower surfaces, 
in contrast to the semicircular leading edge of the 
1903 section and the rounded one of the 1905. This 
deviation from the Wrights' other practice does not 
seem to be characteristic of the 1907 machine from 
which the section was presumably taken (Reference 
42). As I will show in later discussions, this feature 
makes it inadvisable to draw conclusions about the 
1903 and 1905 wings from the Eiffel data on the #10 

The 1903 and 1905 sections have front spars 1.75" 
deep, rear spars 1.25" deep. The 1903 section tapers 
more or less uniformly from just behind the front spar 
to the trailing edge, being just 1.25" deep at the rear 
spar location. As a result the rib is built in two pieces, 
which are connected across the spar by thin metal 
strips (Figure 31d). The 1905 rib is heavier, with l A" 
x Vz" cap strips instead of the l A" x 3 / 8 " of the 1903, 
and its thickness remains at 1.75" until a point aft of 
the rear spar, so that the rib is built in one piece and 
its cap strips span the rear spar. The rib is attached 
to the front spar by metal strips in both instances. 

The two ribs differ in trailing-edge configuration 
also. They accurately outline the airfoil section except 
at the trailing edge, where the ribs have substantial 
thickness, whereas the actual trailing edges are wires 
attached to the ribs. In the case of the 1903 rib, it is 
attached to the upper surface at the trailing edge, 
which gives the mean line a slight upturn over its last 
5%, while the 1905 trailing edge wire is attached to 
the mid-point of the rib trailing edge. The 1903 rib 
has, as a result of this slope reduction, lower values 
of a and c m . 

Existing drawings do not exactly represent the 
Wright contours. Neither rib has in its list of ordinates 
those of the .025, .05, and .95 stations, which must 
be learned from a detailed scale drawing of the rib, 
or from a replica rib. The rib strips are of ash, which 


^m'\^' % ' 








Sssl!:" ' ' 




f mwmmm 

31. Replica wing ribs and details. 

Longitudinal Dynamics 


does not take placidly to being wrestled around 
arbitrary corners, so that it was necessary to make 
some very slight adjustments in the given dimensions 
in certain spots, although these are entirely negligible 
in aerodynamic effects. The 1903 drawings (Reference 
23) give upper and lower ordinates for the 10% 
stations, while the 1905 drawings (Reference 25) give 
lower ordinates only for each 6" station, leaving the 
upper ordinates to be established by the detailed 
dimensions of the rib. 

A lack of sufficient spacing blocks aft of the rear 
spar gives the 1903 rib a considerable degree of 
compliance, 2.3 lb/in as measured directly at the 
trailing edge. It was at first feared this would have an 
appreciable effect on the aerodynamics of the section. 
However, it proved that the amount of the deflection 
was less than had been originally estimated because 
the lift pressures were less at the section trailing edge, 
and the deflections had less than the estimated effect 
because of the peculiar "parallel link" nature of the 
deflection, with less than expected angular deflection 
of the trailing edge (see Figure 37 and Table 6). 

For purposes of integration the ordinates of the 
1903 mean line were interpolated to 80 segments. 
Finite-difference methods were used to minimize 
sudden changes in curvature. These ordinates are 
given in Table 3, and the section, as well as its mean 
line, are shown in Figure 32. The mean-line was 
derived from the section ordinates as follows: 

y = 

where yl = Ul 

L + U L + U 

(1 -x) 

The chord line of this section passes through the 
upper surface instead of the lower surface at the 
trailing edge. The equivalent angle of attack for the 
mean line is a + .47° for equivalent aerodynamics, 
and that amount has been added to bring the section 
to the more conventional configuration. 

. ■ 




= :f 

i S : :.l 



" \ 








-- ■ 



r ; 

>S S-^ 

i ■;■- 

/-— " 









"""sj > 


\ ^\ 


* . 


.8 size 


1 i 


\ h 

! -i 

t . 1 .-:..; 

52. 1903 Flyer wing section. 

Appendix 2 


A. Lift 

Munk's "General Biplane Theory" (Reference 20) 
tabulates the factors that apply to the lift of 2- 
dimensional biplanes as functions of those of 2- 
dimensional monoplanes. These are functions of g/c, 
and for the Wright g/c of .93 he gives B = .84. Taking 
the monoplane lift as CI = 2-7TAb results in a lift- 
curve slope of SI = 1.68 for the 2-dimensional biplane. 
For the 3-dimensional monoplane with elliptical lift 
distribution Glauert (Reference 21) shows that 


Ab, = 


/here A = b 2 /S 

where Ab and AB[ are 2-d 

and 3-d values of Ab 
Dividing through by CI gives 

_L 1 _L 

So ~ SI ttA 

Glauert then provides the factor t to account for the 
rectangular wing for which 

-L_I + -L ( i + „ 

So SI ttA 

and t has a value of .166. However, the rounded wing 
tips of the 1903 Flyer remove it from the category of 
rectangular wings and I have arbitrarily reduced that 
quantity to .12. 

Munk tabulates Equivalent Monoplane Span factors 
for biplanes in terms of g/b, and for the 1903 Flyer's 
g/b of .148 he gives K = 1.3 so that the Equivalent 
Monoplane Aspect Ratio becomes 

A = <■•"»?>•»>' = 4.072 


and SI 




x 1.12 = .277 

1.68-tt 4.072-it 

SI = 3.61 and si = .063 

The 1903 wing ordinates were interpolated to 
provide 80 chordal segments for integration by the 
method of Glauert (Reference 22) to provide angle of 
zero lift (Az = -4.815) and coefficient of pitching 
moment (Cq = - .0997). For comparison purposes 
and for analysis of sections with fewer available 
ordinate values, the method of Pankhurst, as cited by 
Abbott (Reference 25), was also used. Pankhurst lists 
a series of integration constants, for the 10% chordal 
stations plus the .025, .05, and .95 stations, such that 

Frederick J. Hooven 


JA(U + L) = 2 ^Ay 

U + L Vo + Lo 

where y = (1 — x) 

7 2 2 

Ul + LI 
x yl = = yO 

Cq = iB(U + L) = 2 i>y 

x=0 x=0 

By this measure ao = —4.79 and Cq = —.0997 

Glauert's integrations were performed by computer 
summation of the 80 chordal segments. He showed 


A = - yfl (x) dx 


where uO 

Cq = 2u0 - - Az 

M 2 


Theodorsen (Reference 26) separates lift into two 
components, one of which is in proportion to wing 
curvature only and is invariant with angle of attack, 
which he calls the "ideal" lift and which I have called 
the "profile" lift. It operates at a fixed center of 
pressure whose position determines the pitch coeffi- 
cient of the wing. The other component, which I call 
the "induced" lift, is in direct proportion to absolute 
angle of attack, and it operates at the aerodynamic 
center. The angle at which the induced lift is zero is 
called the "ideal" angle of attack, and it is character- 
ized by an absence of flow discontinuities about the 
leading edge of the thin-wing section. Theodorsen 
shows that 

(2) A t = | = yf3(x)dx 

where fl(x) = 

f2(x) = 

and f3(x) = 


R(l - 


1 - 2x 


1 - 2x 

andR = (x(l - x)) 

i/ 2 

2R 3 

The 80 mean-line ordinates along the chord from 
x = .0125 to x = .9875 are summed according to the 
above functions to obtain ao, Cq and differential 
coefficients for lift distributions. At the same time 
actual forces are integrated and with them the deflec- 
tions of the rib trailing edge. From the differential 
lift coefficients and their chordal positions the pitch- 

ing-moment coefficients are summed. Separate sets 
of figures are kept for profile and induced components 
of lift. 

Theodorsen's procedures are used for the compu- 
tation of the differential lift coefficients. He shows 
that for the profile lift 

de 2 

(3) p = 4R — + - [e - eo + (eo - 2e + el)x] 

dx R 

1 f 1 ydx 

wheree(xe) = nJ R(x7"^ 

el = -Ao 
and eO = 2at - el 

Pressure coefficients for induced lift are 


(4) q = — (a - at) 


Although f,, f 2 , f 3 all become infinite at x = and x 
= 1, for ordinary leading edges the integrand ap- 
proaches zero faster than x/c in equations (1) and (2) 
and at the leading edge in equation (2). Since (2) 
tends to infinity at the trailing edge, it is integrated 
only from x = to x = .95, then the function dBo 
is added to the result, assuming that the trailing edge 
and leading edge are of the form 

y = A + Bx + Cx 2 
Accordingly dBo = .964y 95 - .0964 

dx r 

where — is the slope of the mean line at x = 1 
dx c 

Similarly (3) is integrated from x = .05 to x = .95 
with added 

/ dv dv 
dAo = .467 (y. 95 - y. 05 ) - -0472 -p- - ~r~ 

\dxc dxe 


where is the slope of the mean line at x = 


Equation (4) becomes infinite at x = so Theodorsen 

po = 4 — : — , r being the leading edge radius 

{2xlcy h - 

Wing-rib deflections were computed from measured 
compliance of the replica rib, which was 2.3 lb/in as 
applied to the trailing edge. Deflections, as shown in 
Figure 37, with a double exposure showing the 
deflected rib superimposed on the undeflected, show 
relatively little angular deflection, which explains why 
the aerodynamic effects of the deflection were much 
less than originally estimated. Lift coefficient vs. 
angle of attack was affected to a small degree as 

Longitudinal Dynamics 


shown by the dotted lines on Figure 33. Since lift 
and moment were affected almost equally, pitching 
moment was little affected as it is plotted against lift 
coefficient. Aerodynamic effects of deflection were 
computed by changing the ordinates in accordance 
with measured deflections, which are tabulated in 
Table 6. 

Inter-wing forces, resulting from mutual repulsion 
forces, having no resultant, were not given further 

Realistic flight simulation demands more than just 
a lift-curve slope, and theoretical guidance is difficult 
in the construction of realistic-looking wing lift curves. 
Knight and Wentzinger (Reference 27) have in a series 
of tests of biplane configurations using the Clark Y 
section, recorded their lift and drag forces from —39° 
to +90° angles of attack. Their lift curve from —5° 
to + 21° is closely approximated by the following 

CI = Slab - KabP where p = 6 

then dCl/da = SI - pKab?" 1 

Assuming the maximum lift will be at a = 16 (ab = 
20.84) at which point 


= 0, 

from which it is determined that k = 4.012 x 10~ 7 . 
The resulting lift-coefficient curve is plotted on Figure 
33. Above normal flying angles the wing properties 
are those of Knight, et al., of interest in simulating 
the behavior of the canard configuration under stalled 
conditions, because of the Wrights' interest in the 
configuration for its resistance to diving after a stall. 
This subject is discussed at greater length under 
"Concluding Remarks; The Simulated Flights." 

It will be seen that two lines are drawn for the lift 
curve in Figure 33 for angles less than At, one of 
which intersects theO lift line at —4.84, the theoretical 
angle of zero lift. The heavy line curves inward from 
the theoretical straight line to intersect lift at — 3.58. 
This value of a„ has been derived from the study 
"Wind-Tunnel Tests of a l A Scale Model of the 
Wright Brothers 1903 Flyer Airplane" by Bettis and 
Culick (Reference 28), who reported lift only for the 
complete machine with stabilizer. 

McPhail (Reference 29), in his "Textbook Analysis 
of the Longitudinal Characteristics of the 1903 Wright 
Flyer," attributes this to the presence of underwing 
burbles, citing Wood (Reference 30), who remarked 
about such burbles and attributed their effect to an 
effective thickening of the wing section, with reduc- 
tion in mean camber, because of effective reduction 
in the camber of the lower surface of the section. 
There is, however, no way in which this effect can 
be related to angle of attack for accurate computation. 

33. Aerodynamics of 1903 Flyer airfoil section. 

The same reduction of the numerical value of a 
may be seen on other tests of early wing sections with 
high camber and small section thickness, such as the 
Eiffel 10 and the Gottingen 483 (Reference 31), 
whose lift curves are seen in Figure 34. The latter 
shows ao' = —4 where a = —5.29. The effect is 
more pronounced for the Eiffel 10, owing to its sharp- 
edged leading edge, where 


2.85 for ao 


Modification of the mean-line ordinates to repro- 
duce the effect of the under-wing burbles was limited 
to the range of angles between at and ao where Cli 
is negative since it was assumed that no burbles would 
exist at the ideal angle of attack. Camber reduction 
was arranged to take place progressively over the 
range, and to be at its maximum at the section leading 
edge, falling off toward the trailing edge. 

\ 2 

ye = y 

n(l -x) 

a a 

v ap — ao 

.833y(y - x) 

a a 


7 ^% 

-v 7 ^ 

_y / 

' z. 


J .A 





71 - - 


._ Jf *■ _ 


7ti '-"'"' 


/ L,' 

/ // 

-J.-£> Z_ i— 



34. Lift curves, Eiffel 10 and Gottingen 483 section. 


Frederick J. Hooven 

35. Lift pressure distribution, 1903 Flyer airfoil 

where n has been arbitrarily chosen so that Cli = 
— Clp at a = —3.58. The multiplying factor was 
squared to provide a curved line that resembles those 
of test reports rather than a pair of intersecting straight 

The resulting lift and pitching-moment coefficients 
are plotted in Figure 33 and tabulated in Table 5. 
Lift-pressure differentials as plotted in Figure 35 
clearly show the effects of the trailing edge slope 
reduction in the negative pressures seen around the 
95% chord point. 

B. Pitching Moments 

Biplane theory does not provide the basis for a 
consistent theory of biplane pitching moments. While 
Munk (Reference 20) makes the distinction between 
Cli and Clp, he defines their limits differently from 
Theodorsen (who wrote 10 years later) and in terms 
that involve a contradiction. Clp is the lift that occurs 
when cp = .5 and equals 

Clp = 2 sin abB 1 ^ Cli = 2 abB 

However, at aa = ao it is necessary that the two 
components of lift be equal and opposite, and the 
two cannot be equal and be the product of different 
factors, B and B 1 ^. Munk himself does not pursue 
the point and in his summary tables he lists the 2- 
dimensional biplane lift factor as B. 

However, Munk and Glauert agree that pitching 
moments for biplanes are substantially unaltered from 
their monoplane values in their relation to lift coef- 
ficient, although they cannot remain unchanged in 
their relation to angle of attack, which is of interest 
in considerations of longitudinal dynamics. For this 
to be true it follows that the centers of pressure be 
farther to the rear in the biplane than in the mono- 
plane, since the lift coefficients are smaller. Glauert 
(Reference 21, p. 180) remarks that the aerodynamic 
center of the biplane moves forward from x = .25 to 

x = .219. Munk (Reference 20, p. 484) puts this 
point at X — .231 for g/c = .93. 

With the integration procedures of Glauert and 
Theodorsen set up, it was easy to integrate for pitching 
moments. It is useful to integrate section lift differ- 
entials separately from those of induced lift, as well 
as separate lift coefficients. 

Cm = pdx + qdx 
Jo Jo 

Clp acts at a fixed center of pressure and is invariable 
with angle of attack, while Cli acts also at a fixed 
center of pressure but is a function of absolute angle 
of attack. 

Cm = Cl p • cp p + Clj • cpi 

and Cq = Cl p (cp p - cp,) 

and the aerodynamic center is cpi 

Moment summations for the monoplane wing show 
cp, = .2614 and cp p = .406. To provide the changes 
for the biplane wing xp and xq have been transformed 
as follows: 

xp = x(l - kl) + klx' /2 

xq = x(l — k2) + k2x 2 where kl and k2 are 
arbitrarily chosen to change cp p from .406 to .4362 
and cp, from .2614 to .2418. 

These figures are less than theory calls for, but 
more in line with test results such as Knight and 
Wentzinger (Reference 27), Noyes (Reference 39), 
and Munk (Reference 40), who tested biplane cel- 

Moments are plotted in Figure 33. In Figure 35 
are shown pressure distributions, and, for comparison, 
some actual-flight pressure distributions of Crowley 
(Reference 41), showing the rearward shift that takes 
place at low angles of attack due to the above-noted 
shifts in centers of pressure. Note that since Clj is 
negative at angles below a, the effect of moving the 
aerodynamic center forward is actually to move the 
center of pressure farther to the rear. This effect can 
also be seen in the test results cited above, where 
the values of Cq are higher in proportion at the lower 
angles of attack. 

The pitching moment resulting from the stabilizer 
section moment is by no means negligible but is dealt 
with separately in Appendix 2C. 

C. Stabilizer Aerodynamics (1903) 

The biplane stabilizers of all the Wright canards were 
constructed with thin, bendable ribs, connected by a 
control linkage that caused them to bend into a camber 
that increased with increasing stabilizer angle of in- 

Longitudinal Dynamics 


cidence. A full-scale mockup of rib and control linkage 

was built, as shown in Figure 38, for the purpose of 

evaluating this camber/incidence relationship, from 

which it was learned that camber is at -9.88° 

incidence, and that it increases approximately linearly 

as follows: 

Yh = .0028(Ah + 9.88) 

Bending-moment analysis of the ribs gives y/Y as a 
function of x 

yh/Yh = 2.157x 4 
yh/Yh = 

3.451x 3 - 2.588x 2 + 3.882x 
when x = .4763 

These values are tabulated for Yh = .05 in Table 3 
for 80 chord stations. 

Thin-line values for ao are computed by the meth- 
ods of Glauert and Pankhurst respectively, — 101. 9Y 
and — 100.32Y so we write 

-aoh = lOOYh = .28(Ah + 9.88) 

Munk's 2-dimensional biplane lift factor for the 
stabilizer's g/c of .852 is B = .825 giving a biplane 
lift-curve slope of 1.65 = 5.18/rad. For the gap/span 
ratio of 26/144 Munk's EMS = 1.15b h = 13.8 

A = 

13.8 2 



and the shape of the Wright 1903 horizontal stabilizer 
is nearly enough elliptical that I have not applied any 
plan-form correction. So therefore 

1 1 1 

SI 3.6 5.18 


SI = 3.55 si = 


Because of its variable camber the stabilizer will have 
an indefinite number of lift-coefficient curves. To 
plot realistic curves for simulation purposes, beginning 
with the uncambered section, presents certain diffi- 
culties because of the scarcity of empirical information 
concerning tests of sections of this configuration. 
Plotted in Figure 36 are the maximum lift coefficients 
of uncambered sections from the NACA 66-series 
(Reference 31). It can be seen that there is a decrease 
in maximum lift as the section thickness and the 
leading-edge radius are decreased. Theodorsen notes 

















— l 



l J 

36. Lift maxima for uncambered NACA 66 — airfoil 

that the "entrance losses" of wing sections are greater 
with decreasing leading-edge radius, but gives no 
quantitative information (Reference 26, p. 418). 

Noting that the stabilizer section of the 1903 ma- 
chine has a leading-edge spar 5 /s" thick with corners 
rounded to 3/16" radius, the latter being .00239 chord, 
while the leading-edge radius of the 66-006 section is 
.00223, maximum-lift figures for that section are taken 
as a useful guide for estimating maximum stabilizer 
lift. Where the 2-dimensional lift for the 66-006 is .8, 
allowing for biplane effects and 3-dimensional losses, 
the uncambered stabilizer lift maximum has been 
taken to be 0.65. A curve has been plotted accordingly, 
as seen in Figure 36, with a family of lift curves for 
different cambers and their corresponding angles of 
incidence. Up to 16° this curve is represented by 
equation (Reference 5) in which p = 5, a max = 13° 
and k = 4.34 x 10" 7 si = .062. 

The leading-edge configuration suggests that under- 
wing burbles will be seen at all negative angles of 
attack. However, negative angles of attack are not 
encountered during steady-state flight configuration 
in which the eg is forward of the 36% station, so no 
effort has been made to find out what the effects of 
under-wing burbles might be. 

Stabilizer downwash is partially neutralized by the 
fact that the stabilizer span is only 30% of the wing 
span, and since the wing will encounter the upwash 
from the attached vortices, the resultant downwash 
will be that of the bound vortex only. According to 
Millikan (Reference 33), the angle of downwash will 
be Clh/A = .09Clh. Tests by Silverstein, et al. 
(Reference 34) show that at 1.3 semi-spans behind 
the wing's quarter-chord point this angle is decreased 
about 20% from its original value, so that we take 
elevator downwash to be 


= .072Clh 

Flying between and 15 feet altitude, as the Wrights 
did, would reduce this to between and .06. 

Elevator downwash has the effect of reducing wing 
lift without substantially affecting stabilizer moment. 
This effect can be readily produced by a reduction in 
effective area, with a corresponding increase in sta- 
bilizer moment arm. This will approximate 

dCl = Sl^ = 3.61 x .072 = .26Clh 

Consequently a 26% reduction in effective stabilizer 
area, with a 35% increase in effective moment arm 
will reproduce this effect accurately in simulated 

The pitching-moment coefficient of the stabilizer 
section, integrated by the Glauert method is 


Frederick J. Hooven 

Cqh = 


= .000374(A h + 9.88) 

Rather than conduct a complete lift-distribution anal- 
ysis of the stabilizer mean-line, this amount has been 
reduced in the same proportion as the observed 
reduction in the calculated pitching-moment of the 
wing section, approximately .066/. 97, so that the 
pitching-moment coefficient of the stabilizer section 
at zero lift has been taken to be 

C = .00025(ih + 9.88) 

The stabilizer geometry of the 1905 machine is like 
that of the 1903, but one difference between the 1905 
machine and its 1907 successor is that on the 1907 
model the pivot point of the control arm is moved 
forward of the line between the pivot points of the 
two surfaces and its length is increased proportionately 
so that the connecting links between the arm and the 
surfaces are nearly vertical in the normal position. 
This difference is easily seen in Figures 38 and 39, 
the latter of which displays the geometry of the 1907 
stabilizer. The configuration is clear from the drawing, 
Figure 11, and in many of the photographs of the 
machine in flight, such as Figure 12. The basic 
dimensions are taken from a letter dated August 9, 
1908, to Orville from Wilbur while he was in France 
(Reference 44). 

D. Drag 

Bettis and Culick (Reference 28), in their tests of a 
Vfe-scale model of the 1903 Flyer, have provided test 
information on total drag of that machine. Their drag 
figures are for the entire machine, including stabilizer, 
while the coefficients refer to the wing area alone. 
The drag data plotted in Figure 33 is for the machine 
without stabilizer, since for simulation purposes it is 
necessary to carry the stabilizer drag as a separate 
component of the total, because of its variability with 
differing values of ih. The ratio of stabilizer drag to 

38. Stabilizer geometry, 1903 Flyer. 

37. Trailing-edge deflection, 1903 Flyer rib. 

39. Stabilizer geometry, 1907 Flyer. 

the total has been taken as equal to the ratio of its 
surface area to the total, which is .103 so the drag 
coefficients of Bettis and Culick run #1, taken with 
ih = 0, have been multiplied by .897 to obtain the 
drag for the machine without stabilizer. 

Stabilizer drag is then reckoned for its unique angle 
of attack, using the same tables as tabulated for the 
rest of the machine, but multiplied by the ratio of 
stabilizer area to wing area. This method, while it is 
an approximation, gives results that are satisfactorily 
close to those given by Bettis and Culick for ih = 
10, IVz and 15, as well as those computed by more 

Longitudinal Dynamics 


exact assumptions concerning stabilizer profile and 
induced drag. 

Drag has a significant effect on longitudinal statics 
in the 1903 Flyer, because it is relatively large in 
proportion to other aerodynamic forces, and because 
the thrust axis is high enough above the eg that thrust 
forces result in a significant pitching moment. Stabi- 
lizer drag also has a pitching-moment component at 
high and low angles of attack. 

Bettis and Culick drag measurements are signifi- 
cantly higher than the Wrights estimated drag. In 
their notebook K (Reference 35) the Wrights' estimate 
was 90 lb, but this was at an estimated forward speed 
of 24 mph. At the Flyer's historic speed of 30 mph 
(44 fps), the drag according to Bettis and Culick would 
be 125 lb and, according to the Wrights' implied drag 
coefficient, it would be 140 lb. The Wrights had 
based this early estimate on an estimated weight of 
630 lb and raised it to 100 lb when they realized that 
the weight would be over 700 lb (see Reference 24). 
They regarded 100 lb as very conservative, and were 
reassured by their measurement of 132 lb static thrust 
after the machine was assembled. Consequently, the 
machine requires 13.8 hp for level flight at 30 mph 
(44 fps) which means that their engine must have 
turned out more horsepower than the Wrights' tests 
had shown. 

The required modifications of historically accepted 
numbers for engine power and propeller efficiency 
can be made without difficulty. Four factors must be 
taken in account: (1) Engine power, with cooler air 
of higher density, would be higher than static tests 
indicated; (2) engine cooling, on a cold day in flight, 
would be much more effective than in static tests, 
and the engine's output was highly dependent on 
cooling, because of the unjacketed valve cages and 
cylinder heads, as witness the decline in power as the 
engine warms up (see Reference 26); (3) induced drag 
was less than wind-tunnel measurements because of 
ground effects at the very low altitudes at which the 
flights took place; and (4) propeller efficiency was 
obviously higher than the Wrights' estimate of 66%, 
which included an exaggerated transmission loss as 
the result of Chanute's misinformation (see Reference 
24) to the effect that chain losses would be 25-30%. 
The Wrights figured 10-15%, while the true losses 
were probably closer to 5%. 

Drag for high angles of attack has been estimated 
from the data of Knight and Wentzinger (Reference 


Longitudinal Statics and Dynamics 

The mechanics of simulated flight will be presented 
in the way they are processed in the computer. 

Aircraft measurements are given as follows for the 
1903 Flyer, x coordinates given in units of c/100 from 
wing leading edge rearward, z coordinates from the 
eg upward. 

1. R = 9.67 (Ratio of wing to stabilizer surface 

2. Dl = —112 (Distance, stabilizer Vi chord to 
wing leading edge) 

3. Wt = 750 (Weight) 

4. Im = 280 (Longitudinal moment of inertia, si 

5. S = 510 (Wing surface area) 

6. ch = 6.5 (Wing chord, ft) 

7. zl = +23 (Vertical coordinate of thrust vector 
upward from eg) 

8. z2 = + 12 (Vertical coordinate of drag center) 

9. z3 = +15 (Vertical coordinate of lift center) 

10. zh = (eg + Dl) sin ap (Vertical coordinate of 
center of stabilizer drag) 

11. Fd - .00125 (Density of air, sl/ft 3 at 40°F) 

Flight conditions supplied by the operator for each 
flight: aa, angle of attack, Z, altitude. 

The data bank is available that consists of a complete 
matrix of four flight coefficients for all angles from 
- 19° to + 30° angle of attack, CI, Clh, Cd, and Cm 
(see Table 5). 

The static conditions for equilibrium in level flight 
are set up according to the specified flight conditions 
in a "takeoff routine. If equilibrium is not possible 
under the conditions specified the flight terminates 

The process flight variables are set up as follows: 

The specified angle of attack is divided into integral 
and fractional parts for data look up and interpolation 
to establish flight variables. CI, Cd, and Cm are then 
looked up for the specified angle of attack and C is 

C = Cr 

Cl(Cg — zl sin a p ) 

To establish equilibrium stabilizer angle, the range 
of angles is scanned by 1° increments as Kh is looked 
up for that angle and corrected for camber according 
to incidence setting. 

Kh - Kl = .017 J/R 

KI is looked up as a function of I + J. When the 
total positive moment exceeds the total negative 
moment, the scan goes back a degree and then 
advances by increments of .01° until equilibrium is 
reached when 

(Dl + Cg)(Kh + Kd sin (AP) 

= C + .25J/R + Kdz2 + Cd(z2 - z3) 


Frederick J. Hooven 

After the stabilizer scan the following variables are 


V = 2Wt(Cl + Kh)/Pd 1/2 

Th = V- pd(Cd + Kd) 

H = Vth/400 

The takeoff routine is completed by a printout of the 
flight variable headings and a list of the variables as 
they exist at takeoff 

T V v 





V B-10 

Flight data is listed in the same columns during the 
flight routine. 

Figure 40 is a force vector diagram of the 1903 
Flyer, and Figure 14 is a plot of all possible values of 
stabilizer incidence vs. angle of attack for equilibrium 
over a range of eg positions from 8 to 35. Regions of 
positive slope are regions of stable flight. 

The flight routine begins with the listing of the 
various optional subroutines, for automatic landing, 
automatic pilot, random wind gusts. Then the stabi- 
lizer angle is corrected for pitch motion 

Be = B - 57.3TAN" 1 (Dl + Cg) ch/100 w/V) 

The following functions are updated by look up, Cd, 
Cm, Kh, Kd, CI 

then Pd = V 2 Sda/2 

L = Pd(Cl + Kh) 

Dg = Pd(Cd + Kd) 

C = Cm - Cl(Cg - zl sin (Ap)) 

gt(L - Wt cos (af)) 

af = 


P = ch/100 (-z2Th + Pd(Cdz3 - C 
+ Kh (Dl + Cg) + Kd sin (Ap))) 


W = W + 

I Im 

Ap = Ap + WT 

Aa = Ap + Ai - Af 

Th = 400 H/V 

V = V + gT[(Th - Dg)/Wt - sin(Af)] 
Vv = VsinAf 

Z = Z + VvT 

40. Vector diagram, 1903 Flyer. 


The Simulation Program 

The preceding discussion shows the nature of the 
processing of the program, basically a straightforward 
interaction of static equations. 

The program is listed in Appendix 1, and is anno- 
tated for those who wish to follow it in detail. Its 
nomenclature is the same as that of the present 

These two flight routines take up pages 3 and 4 of 
8, of which 1 is nomenclature, 2 is data assembly, 5 
is landing and printout routines, 6 is devoted to control 
options, 7 to subroutines and 8 to data. 

The operator, who specifies cg location, angle of 
attack, and altitude before the flight takes off, controls 
stabilizer incidence and throttle, although manual 
stabilizer control is not possible in unstable flight 
regimes. Throttle is specified in terms of power. Any 
amount of power is available. For unstable flight an 
automatic pilot is necessary. Its reaction time is 
controllable between and .08 seconds, and its 
sensitivity can also be adjusted. 

As page 6 of the program will show, there are many 
options available to the operator who may vary the 
frequency of printouts and pauses for instructions, to 
begin and end wind-gust routines, to activate auto- 
matic pilots and lander, or to eliminate further instruc- 

Longitudinal Dynamics 




Aircraft dimensions are measured along longitudinal axis 
in units of percent wingchord from leading edge rearward. 
All motion is in the vertical longitudinal plane. 


Ai,Aa,Ap,Af= Angles of incidence, attack, pitch and path, radians 

B = Elevator-to-wing angle, degrees +10 

Bl = Effective angle of elevator as corrected for pitch motion 

B3= Automatic pilot time-delay index 

C = Pitching-moment coefficient about center of gravity 

Cg = Position coordinate of airplane center of gravity, percent chord 

ch = Wing chord, feet 

Cm = Pitching-moment coefficient 

Cd = Drag coefficient for complete aircraft except elevator 

Dg = Total drag of airplane less stabilizer, pounds 

DL = Distance from leading edge to aerodynamic center of elevator 

F(I), B3 = Pilot-reaction-time factors 

Fd = One half density of air times wing surface area 

Kl,K(I) = Elevator lift coefficient as applied and as tabulated 

GV,V9 = Velocity values for random wind gusts 

H = Horsepower at engine shaft, assuming propulsion efficiency of 72.7% 

1, 12 = Angle of attack integral degrees +10 and hundredths 

Im = Aircraft moment of inertia, slug-feet squared 

J,J2 = Elevator angle integral degrees +10 and hundredths 

Kh = Lift coefficient of elevator corrected for variable camber 

Kd, E(J) = Drag coefficient of stabilizer 

K1,K(I) = Elevator lift coefficient 

Lz = Lift of wing and elevator, pounds 

Ml— -Mn —Flag variables 

N = Time increment count 

P = Pitching moment 

Pd = Dynamic pressure, lbs/sq.ft. 

Q = Input variable for instructions 

Ql, Q2, Q3 = Angle-of-attack, sensitivity and flag for automatic pilot 

Q4 = Autopilot damping factor 

R= Ratio of wing area to elevator area 

S = Wing surface area, square feet 

T= Finite time increment, seconds 

Th = Thrust, pounds 

T8,T9 = Random-wind gust duration, seconds 

U,U1 = Integral and fractional angle of attack of elevator 

V = Airspeed, ft/sec 

Vv = Vertical velocity 

W = Rate of change of pitch angle 

Wt = Weight of aircraft, pounds 

XI = Horizontal distance flown, feet 

Z = Altitude, feet 

Z1,Z2,Z3 = Vertical coordinates of center of life, drag and thrust 


In the following routine the specifications and aerodynamics 

of the Flyer are read into the program. 
2120 LET R = 9.7 (Dimensions of the Flyer are given) 
2140 LETDL = 112 
2160 LETWt = 750 
2180 LETIm = 280 
2200 LETS = 510 
2220 LETch = 6.5 
2260 LETZ1 = 15 
2280 LETZ2 = 23 
2300 LETZ3=13 
2320 LETCg = 29 

2340 LET Q2,Q4= 1.5 (Autopilot settings) 
2360 LETB7 = 2 
2380 LETB3 = 3 

72 Frederick J. Hooven 

2440 LET Ai = 8/57.3 (Angle of incidence) 

2480! In this section the aerodynamic data is read out of the data bank 

2500! and certain variables are given initial values. 

2540 DIM D(110),E(110),F(10),G(110),K(110),L(110),M(110) 

2560 FOR I = 1 TO 90 (Lift coefficients) 

2580 IF K41 THEN 2640 

2600 LET Lz=720-8*I (This algorithm for approximate coefficients from 30--80°) 

2620 GO TO 2660 

2640 READ Lz 

2660 LETL(I) = Lz/1000 

2680 NEXT I 

2700 FOR I = 1 TO 1 10 (Stabilizer Lift) 

2710 IF K51 THEN 2740 

2720 LET Kl =550-5*1 

2730 GO TO 2750 

2740 READ Kl 

2750 LETK(I) = K1/1000/R 

2760 NEXT I 

2770 FOR I = 1 to 1 10 (Drag coefficients) 

2780 IF K51 THEN 2810 

2790 LET Cd = 50+ 10*1 

2800 GO TO 2820 

2810 READ Cd 

2820 LET D(I) = Cd/1000 

2830 NEXT I 

2840 FOR 1=1 to 90 (Pitching-moment coefficients) 

2850 IF K41 THEN 2880 

2860 LET Cm = 140 + 4*1 

2870 GO TO 2890 

2880 READ Cm 

2890 LETM(I) = Cm/10 

2900 NEXT I 

2910 LET T= .02 (The time period of iteration) 

2920 LET Fd = .00125*S ( 1/2 density of air and wing surface area) 

2940 LET Af,M4,M6,M7,M8,M9,N,N9,Ql,Q3,Vv,W,Xl =0 (Reset for a new flight) 


2970 PRINT "Z,Aa"; 

2980 INPUT Z,Aa (Operator inputs altitude and angle of attack) 

2990 LET Ql = Aa (This sets the automatic pilot target angle) 

3000! Next is the takeoff routine in which are established the variables 

3020! for steady-state flight under the conditions specified in the data input. 

3040 FOR 1= 1 TO B3 (Reset the time-lag loop for the autopilot) 

3048 LETF(I) = 

3054 NEXT I 

3060 LET I = INT(Aa) + 10 (Set the data matrix in line with the angle of attack 

3080 LET I2 = Aa+ 10 — I including the fractional value for interpolation) 

3100 LET C1 = L(I) + I2*(L(I + 1)-L(I)) (Look up and interpolate for lift coefficient) 

3120 LET Aa = Aa/57.3 (Change angle of attack to radians) 

3140 LETCd = D(I + 10) + I2*(D(I + ll)-D(I + 10)) (Lookup and interpolate for drag coefficient) 

3160 LET Cm = M(I) + I2*(M(I + l)-M(I)) (Lookup and interpolate for pitching-moment coefficient) 

3180 LET Ap = Aa-Ai (Set pitch angle for level flight) 

3200 LET C = Cm-Cl*(Cg-Zl*SIN(Ap)) (Compute C including "pendulum" pitch moment) 

3220 PRINT "C = ";C, 


3240 FOR J = 1 TO 37-1 (Scan the range of stabilizer angles) 

3280 LET K(J) = K(I+J) + I2*(K(I+J+1)-K(I+J))+.17*J/R (lookup stabilizer lift and for stabilizer drag) 

3300 LETE(J) = (D(I+J) + I2*(D(I+J + 1)-D(I+J)))/R 

3320 IF (DL + Cg)*(K(J) + E(J)*SIN(Ap))>C + .25*J/R + E(J)*Z2 + Cd*(Z2-Z3) THEN 3480 

3400 NEXT J (When the stabilizer lift equals the pitch moment equilibrium has arrived) 

3420 PRINT "NO TAKEOFF" (If the lift is not enough the flight terminates) 

3440 GO TO 2940 


3480 LET J=J — 1 (Set the angle back 1 and scan .01 degree at a time) 


3520 FOR J2 = TO .99 STEP .01 

3540 LETKh = K(J) + J2*(K(J + l)-K(J)) 

3560 LETKd = E(J) + J2*(E(J + l)-E(J)) 

Longitudinal Dynamics n ~i 

3580 IF (DL + Cg)*(Kh + Kd*SIN(Ap))>C + .25*(J + J2)/R + Kd*Z2 + Cd*(Z2 -Z3) THEN 3640 

3600 NEXT }Z 


3640 LETB.=J+J2 

3720 LET V, Vh = SQR(Wt/Fd/(Cl + Kh)) (There are separate Vs for stabilizer and wing for gusts) 

3760 LET Th = V*V*Fd*(Cd + Kd) 

3780 PRINT "V = ";V, 

3800 LETH = Th*V/400 

3820 PRINT INT (100*H)/100, 

3840 LET Ml = 1 (This is the "in flight" flag) 

3860 PRINT "T Vv Z Aa Ap Af W XI V B" 

3880 LET Pd = V*V*Fd (Print out flight data headings) 

3940 GO TO 5600 (Go to print routine to print out initial flight variables) 

3980 PRINT 

3990 LET V5 = V4 (This is part of a routine for flagging when Vv goes through 

3999 LET V4 = Vv zero, through a maximum, or through a minimum) 

4000! This is the Regular Flight Routine. 

4020 IF Q3 = 1 THEN 7420 (Flag for the automatic4anding subroutine) 

4040 IF M6 = THEN 4100 (If it is up the landing begins when we are 1 1/2 

4060 IF -Vv>2*Z/3 THEN 7360 seconds from ground) 

4100 IF N9 = THEN 4140 (Flag for the random-wind-gust subroutine) 

4120 IF N = >T9 THEN 7820 (If it is up we wait for the gust duration to end) 

4140 IF Q1<>0 THEN 7525 (Flag for the automatic-pilot subroutine) 

4180 LET Bl =B-57.3*ATN((DL + Cg)*ch/100*W/V) (Correct stabilizer angle for pitch motion) 

4200 LET I = INT(57.3*Aa) + 10 (Set matrix to angle of attack) 

4220 LET I2 = 57.3*Aa+10-I 

4240 LET U = INT(I + I2 + B1) (Also to stabilizer angle of attack) 

4260 LET Ul =1 + 12 + B1-U 

4300 LET Cd = D(I + 10) + I2*(D(I + 1 1) - D(I + 10)) (Lookup for drag coefficient) 

4310 LET Cm = M(I) + I2*(M(I+l)-M(I)) (Pitching-moment coefficient) 

4320 LET Kh = K(U) + U1*(K(U + 1) - K(U)) + .017*B/R (Stabilizer lift) 

4330 LET Kd = (D(U) + U1*(D(U + 1 ) - D(U)))/R (Stabilizer drag) 

4340 LET CI = L(I) + I2*(L(I + 1) - L(I)) (Lift coefficient) 

4342 IF NOT7 THEN 4350 (Time a wind gust from stabilizer to wing) 

4344 LET V = V + V9*N9 (Change airspeed by the amount of the gust) 

4350 LET Pd = V*V*Fd (Dynamic pressure at wing) 

4355 LET Ph = Vh*Vh*Fd (Dynamic pressure at stabilizer) 

4360 LET Lz = Cl*Pd + Kh*Ph (Compute lift force and drag force) 

4380 LET Dg = Cd*Pd + Kd*Ph 

4400 LET C = Cm-Cl*(Cg-Zl*SIN(Ap)) + .25*B/R (Pitching moment coefficient about eg) 

4420 LET Af = Af + T*32*(Lz - Wt*COS(Af))AVt/V (Correct flight path bv applied forces) 

4440 LET P = ch/100*(-Z2*Th + Pd*(Cd*Z3-C) + Ph*((DL + Cg)*Kh + Kd*SIN(Ap))) (Pitching moment) 

4460 LET W = W + P/Im*T (Pitching-motion angular velocity) 

4480 LET Ap = Ap + W*T (Pitching angle updated) 

4500 LET Aa = Ap + Ai-Af (Angle of attack updated) 

4520 IF Aa*57.3<50 THEN 4580 


4560 GO TO 5260 

4580 IF Aa*57.3> - 7 THEN 4620 


4610 GO TO 5160 

4620 IF U + UK110 THEN 4650 

4630 PRINT "ELEVATOR STALL, U + U1 = ";U + U1 

4640 GO TO 5160 

4650 IF U + Ul >2 THEN 4680 


4660 GO TO 5160 

4680 LET Th =400*H/V (Compute thrust) 

4720 LET DV = 32*T*(SIN(Af) + (Dg - Th)/Wt) (Update velocity) 

4730 LETV = V-DV 

4740 LETVh = Vh-DV 

4760 LET XI =X1 +V*T*COS(Af) (Update horizontal distance flown) 

4840 LET Vv = V*SIN(Ap) (Compute vertical velocity) 

4860 LET Z = Z + Vv*T (Update altitude) 

4880 LET N = N + 1 (Keep track of time) 

4960 IF Z>0 THEN 5480 (If there is still altitude, continue flight with printout decision.) 

5000! This is the landing routine 

74 Frederick J. Hooven 

5020! If the rate of descent is less than 6 ft/sec when the 

5040! aircraft reaches the ground, it's a good landing. 

5060! If it's between 6 and 10. ft/sec it's a hard landing. 

5080! If it's more than 10 it's a crash. 


5120 IF Vv> - 8 THEN 5200 

5140 IF Vv> - 12 THEN 5240 

5160 PRINT "Crash at"; 

5180 GO TO 5260 

5200 PRINT "Land at"; 

5260 LET Ml =0 

5280 PRINT N*T; "Sec" 

5290 LET X = X + V*(INT(N/50) + l -N/50) 

5300 GO TO 5600 


5320! This is the printout routine. 

5340! Printout normally takes place one per second of flight time. 

5360! Printout also takes place at maximum and minimum values of the 

5380! vertical velocity, Vv. 

5440! By means of special instructions printout may be caused to 

5460! take place at shorter intervals. 


5480 IF V4*Vv*(V4- V5)*(V4- Vv)>0 THEN 5600 (Prints if Vu goes through zero, maximum or minimum) 

5500 IF M8>N THEN 5600 (Flag for Vsoth-second printout) 

5520 IF INT(N/5)<N/5 THEN 3990 (Sieve for '/ioth-second printout) 

5540 IF M9>N THEN 5600 (Flag for Vioth-second printout) 

5560 IF ABS(W)>20/57.3 THEN 5600 (If pitching motion gets over 207sec, print each '/,oth sec.) 

5580 IF INT(N/50)<N/50 THEN 3990 (Sieve for 1-second printout) 


5620 PRINT N*T;TAB(7); 

5640 PRINT INT(100*Vv + .5)/100;TAB(14); 

5645 IF Z<50 THEN 5660 

5650 PRINT INT(Z + .5);TAB(21); 

5655 GO TO 5680 

5660 PRINT INT(100*Z + .5)/100;TAB(21); 

5680 PRINT .01* INT(5730*Aa.5);TAB(28); 

5700 PRINT . l*INT(573*Ap - .5 - 10*Ai);TAB(35); 

5720 PRINT INT(573*Af)/10;TAB(42); 

5740 PRINT INT(5730*W + .5)/100;TAB(49); 

5760 PRINT INT(Xl); 

5780 PRINT TAB(55); 

5800 PRINT INT(10*V + .5)/10; 

5820 PRINT TAB(62); 

5860 PRINT INT(100*(B- 10) + .5/100; 

5880 PRINT INT(100*F(B3))' 

5960 IF Ml = 1 THEN 6200 (If the "in flight" flag is still up go on with the flight) 

5980 GO TO 2940 

6000! This section contains the routines for special instructions. 

6020! Normally the program will ask for instructions one per second 

6040! at the time of printout. This may, at option, be changed to occur 

6060! ten times per second, once every ten seconds, or not at all. 


6100! In-flight changes can be made in elevator angle and throttle setting; in 

6120! windspeed for simulating wind gusts. An angle of attack can be entered 

6140! that will activate the autopilot, causing it to control the elevator 

6160! to maintain that angle. An automatic landing sequence may be activated. 


6200 IF INT(N/5)<N/5 THEN 3980 (These are sieves for '/ioth and 1 second instruction pauses) 

6220 IF M7>N THEN 6320 (Flag for .1 second instruction interval) 

6240 IF INT(N/50)<(N/50) THEN 3980 

6260 IF M4 = 2 THEN 3980 (Flag for no instructions at all) 

6280 IF N<M4 THEN 3980 


6320 INPUT Q 

6360 IF Q>499 THEN 6800! Enter 500 + 10*Q1, the autopilot angle of attack 

6380 IF Q>299 THEN 7180! ENTERS 300 plus a time without further instructions 

6400 IF Q>99 THEN 6740! Enter 100 plus desired horsepower 

Longitudinal Dynamics 75 

6420 IF Q<23 THEN 6660! Numbers less than 22 command elevator angies 

6440 IF Q = 23 THEN 7100! 23 activates the landing system 

6480 IF Q = 24 THEN 6700! Complete the flight without further instructions 

6500 IF Q = 25 THEN 6940! Gives .1 sec instruction pause, with printout 

6520 IF Q = 26 THEN 7020! Print out every '/so second for 1 second 

6540 IF Q = 27 THEN 7060! Print out every .1 second for 1 second 

6560 IF Q = 28 THEN 6900! Print out and instructions each second 

6580 IF Q = 29 THEN 5980! 29 signals to terminate the flight 

6600 IF Q = 30 THEN 6960! Q = 30 Cancels the wind-gust routine 

6620 IF Q>30 THEN 6840! Enter 30 + random wind-gust velocity 

6660 LETB = Q+10! - 10<Q<23 calls for an elevator setting 

6680 GO TO 7200 

6700 LETM4 = 2 

6720 GO TO 7200 

6740 LETH = Q-100 

6750 GO TO 7200 

6800 LETQl = .l*(Q-500) 

6820 GO TO 7200 

6840 LET N9 = 1 

6860 LETGV = Q-30 

6880 GO TO 7200 

6900 LETM4,M7,M8,M9 = 

6920 GO TO 7200 

6940 LETM7,M9 = 49 + N 

6960 LETN9 = 

6980 GO TO 3990 





7000 GO TO 7200 

7020 LETM8 = N + 50 

7040 GO TO 7200 

7060 LET M9 = N + 50 

7080 GO TO 7200 

7100 LETM6=1 

7120 GO TO 7200 

7180 LETM4 = 50*(Q-300) + N 

7200 IF INT(100*Q + .1)/10 = INT(10*Q + .01) THEN 3990 

7220 LETQ = INT(10*Q)/10 

7240 GO TO 6300 

7320! The final section of the program contains the various subroutines 

7340! The landing subroutine follows 

7360 LETQ3 = 1 

7370 LET Q2,Q4 = 2 (Sharpen up the autopilot for the landing) 


7420 LET B3 = 1 

7520 LETQl=6 + V/20-.l*Z-.l*Vv(Set the landing speed target) 

7525! The autopilot routine follows 

7528 IF ABS(Q1 -57.3*Aa)>l THEN 7535 (This is 1 -degree offset for angle of 

attack response) 

7530 LET F(B3) = 1.5*Q2*(Ql/57.3-Aa/2-(Ai + Ap)/2) 

7532 GO TO 7538 

7535 LET F(B3) = 1.5*Q2*(Ql/57- Aa) 

7538 IF ABS(W)<.01*B7 THEN 7544 (This is offset for pitch-damping response) 

7540 LETF(B3) = F(B3)-1.5*Q4*2-.00014*P 

7544 FOR I = 1 TO B3 - 1 (This is the time-lag loop) 

7548 LET F(I) = F(I + 1) 

7552 NEXT I 

7556 LET B = B + F(1) (The autopilot updates the stabilizer control) 

7580 IF U + UK35 THEN 7640 (Don't move the control past the stall point) 

7600 IF Q2*F(B3)<0 THEN 7640 

7620 LETF(B3) = 

7660 IF U + U1>10 THEN 7720 (Or into a negative-lift situation) 

7680 IF Q2*F(B3)>0 THEN 7720 

7700 LETF(B3) = 

76 Frederick J. Hooven 

7720 GO TO 4180 

7820! The random-wind subroutine follows 

7830 LET Vh = V (The maximum gust speed is specified) 

7840 IF N9<0 THEN 7900 

7860 LET V9= RND*(GV — 0) (Gust speed is random amount up to that maximum) 

7865 LETGA = RND*(GV/500-0) 

7870 LET A5 = A5 + GA*N9 

7880 LET T8 = INT(RND*(150-20)) + N (Gust timing is random up to 3 seconds) 

7900 LET Vh = Vh + V9*N9 

7920 LET T7 = INT(450/Vh) + N (Times the gust from stabilizer to wing) 

7940 LETN9=-N9 


7970 LETT9 = T8 + N 

7980 GO TO 4140 

8000! The aerodynamic data bank follows 


8160! L(I) Wing lift coefficients in absolute units *1000 


8200 DATA -270, -248,-208,-168,-128,-88,52,142,225,300! -9,0 

8220 DATA 367,430,493,556,618,680,740,800,857,912! 1,10 

8240 DATA 963,1010,1050,1082,1103,1110,1110,1072,1017,932! 11,20 

8260 DATA 810,760,730,700,700,700,700,700,705,710! 21,30 


8300! K(I) Lift coefficients of elevator section without camber 


8340 DATA -500,-500,-500,-528,-590,-624,-634,-624,-602,-567! -19,-10 

8360 DATA -523,-474,-419,-363,-304, -244,-183,-122,-61,0! -9,0 

8380 DATA 61,122,183,244,304,363,419,474,523,567! 1,10 

8400 DATA 602,625,634,624,590,528,510,500,500,500! 10,20 

8420 DATA 490,460,440,420,400,380,360,340,320,300! 21,30 


8460! D(I) Coefficients of drag for total aircraft less elevator 


8500 DATA 387,376,353,332,301,272,245,217,190,168! -19,-10 

8520 DATA 152,136,122,109,100,91,84,77,75,73! -9,0 

8540 DATA 75,80,84,91,100,109,122,136,152,168! 1,10 

8560 DATA 190,217,245,272,301,331,353,376,387,399! 11,20 

8580 DATA 415,430,443,457,468,480,493,507,518,530! 21,30 


8620! M, Pitching-moment coefficients 


8660 DATA 34,34,34,34,34,54,72,101,127,148! -9,0 

8680 DATA 165,180,195,210,224,239,254,269,284,299! 1,10 

8700 DATA 313,328,337,344,350,353,350,345,338,334! 11,20 

8720 DATA 319,308,297,286,286,286,286,286,286,286! 21,30 


99999 END 

Longitudinal Dynamics ' ' 

Propulsion Systems 
of the Wright Brothers 



Everything concerning the Wright brothers and their 
achievements has been said before, by the Wrights 
themselves and by later students of their work. 
Consequently, this paper affords a summary of the 
Wrights' propulsion work that was so essential for the 
successful flight of their airplanes. The reader inter- 
ested in an expanded knowledge of the Wrights' 
propulsion systems should consult References 1 through 

From their earliest investigations, the Wright broth- 
ers thought in terms of powered flight. But first they 
realized that the key elements of aircraft control and 
configuration had to be determined and demonstrated. 
As practical engineers, they chose to pursue their 
work in a step-by-step procedure, using gliders with- 
out the complication of a propulsion system. By the 
end of 1902, they had solved the problem of flight 
control, had determined a practical basic configuration 
for the airplane, and had acquired enough actual flight 
experience to feel comfortable in handling a powered 
machine. In the meanwhile they collected information 
on propellers and engines and thought seriously upon 
the matter. 

To propel their airplanes through the air, the 
Wrights knew that they had to have thrust. They had 
to develop a propulsion system. In those days thrust 
was produced by a propeller. In order to produce 
thrust, the propeller required energy to turn it. To 
develop energy, an engine was required. However, 
to run the engine properly, a support system was 
needed to supply fuel, oil, and cooling water. All 
three systems — thrust, energy, and support — were 

required to provide propulsion. Lacking any one, 
nothing would happen. As elements of these systems 
were not readily available, the Wrights had to develop 
all three. 

Probably the most difficult of the three to develop 
was the propeller. At that time the Wrights had no 
real understanding of the air propeller — how it worked, 
how it should be designed. There was no valid 
propeller theory in their knowledge. 

Frank W. Caldwell, the eminent propeller designer, 
pointed out that 

Some time before any actual flights in airplanes, a 
theory of the air propeller was evolved more or less 
independently by Lanchester, Drzweiecke, and 
Prandtl in Germany. This theory is based on the 
conception of the airplane propeller as a series of 
wing sections moving in a spiral path. Knowing the 
characteristics of these wing sections, it was possible 
to predict the performance of the propeller by means 
of mathematical calculations. It is interesting that the 
Wright brothers evolved this theory independently 
without knowledge of the work of the earlier 
scientists. 1 

The problem facing the Wrights, to design a suitable 
and efficient propeller, was amply stated by them in 
an article in The Century Magazine, September 1908: 

Our table made the designing of the wings an easy 
matter; and as screw propellers are simply wings 
traveling in a spiral course, we anticipated no trouble 
from this source. We had thought of getting the 
theory of the screw propeller from the marine 
engineers, and then, by applying our tables of air 
pressure to their formulas, of designing air propellers 
suitable for our purpose. But so far as we could learn, 


the marine engineers possessed only empirical 
formulas, and the exact action of the screw propeller, 
after a century of use, was still very obscure. - 

Further, Orville Wright gave a graphic description 
of the perplexing problem they had to solve: 

It was apparent that a propeller was simply an 
aeroplane (aerofoil) traveling in a spiral course. As we 
could calculate the effect of an aeroplane traveling in 
a straight course, why should we not be able to 
calculate the effect of one traveling in a spiral course? 
At first glance this does not appear difficult, but on 
further consideration it is hard to find even a point 
from which to make a start; for nothing about a 
propeller, or the medium in which it acts, stands still 
for a moment. The thrust depends upon the speed 
and the angle at which the blade strikes the air; the 
angle at which the blade strikes the air depends upon 
the speed at which the propeller is turning, the 
speed the machine is traveling forward, and the 
speed at which the air is slipping backward; the slip 
of the air backward depends upon the thrust exerted 
by the propeller, and the amount of air acted upon. 
When any of these changes, it changes all the rest, as 
they are all interdependent upon one another. But 
these are only a few of the many factors that must be 
considered and determined in calculating and 
designing propellers. 3 

To solve this problem, the Wrights rejected a long 
series of experiments to devise a suitable propeller 
by trial-and-error methods as too costly in time. They 
elected to develop a theory from which they could 
calculate the propeller performance and dimensions. 
The Wrights became so obsessed with the problem 
that they spent long hours in study and debate to 
develop an understanding of the complex subject. By 
February 1903, they had progressed to the point 
where the first full-sized propeller was made and 
statically tested. The propeller had a diameter of 8V2 
feet, a blade width of 6 inches, pitch of 15 degrees, 
and gave 18% pounds of thrust at 245 revolutions per 
minute. Afterward they came to realize that "the 
thrust generated by a propeller when standing sta- 
tionary was no indication of the thrust when in 
motion." 4 

Their propeller theory emerged gradually with their 
formulae improving with experience. Unfortunately, 
the Wrights never organized their theory in a logically 
written form that others could understand. What is 
known is extracted from various entries of formulae, 
test results, and tables scattered among their many 
notebooks. Using these sources, McFarland, in Ap- 
pendix III of The Papers of Wilbur and Orville Wright, 
assembled as good a review as possible of their theory. 

From their propeller performance diagrams, the 
Wrights developed formulae for thrust, torque, slip, 
blade width of an ideal screw, and others. Using the 
formulae, they then developed tables of quantities 
from which they could choose the propeller charac- 
teristics to fulfill their needs. 

Modern propeller theory is based upon two theories: 
the blade-element theory and the momentum theory. 
The two theories complement each other. The theory 
developed by the Wrights combined features of both 
theories. 5 

At first, and through 1904, the blade element used 
was an arc. Wright aerofoil no. 9 was the most efficient 
arc tested in the wind tunnel experiments and un- 
doubtedly was the surface employed. In 1905 the 
propellers were redesigned to produce greater thrust 
and efficiency and given a new shape. The blade 
element in the redesigned propeller was a parabolic 
curve of Wright aerofoil no. 12. While these profiles 
were not very efficient for low incidence, they were 
quite good for high angles and, consequently, good 
for take-off. 

On the application of their propeller to the airplane, 
the Wrights made three basic decisions: 

1. Two propellers would be used, each absorbing half 
the powerof the engine and producing half the needed 
thrust. Two propellers would give a reaction against 
a greater quantity of air with larger pitch angle than 
was possible with a single propeller. Consequently, 
the propeller rotational speed could be lower. 

2. The propellers, to eliminate torque effects upon 
control of the airplane, would turn in opposite direc- 
tions, thereby cancelling the torque effect of each 

3. The propellers would be mounted at the rear of 
the wings as pushers to eliminate the effect upon the 
wings of turbulent airflow over them from tractor 

Another important decision had to be made: how 
to drive the propellers. Little was known of desirable 
or limiting speed ratios between engines and propeller. 
Should it be a direct-driven single propeller attached 
to the engine's crankshaft? Or, should it be remotely 
mounted propellers driven indirectly by the engine? 
The direct-driven propeller offered no flexibility to 
vary experimentally engine and propeller speed ratios 
while providing an undersirable out-of-balance torque 
force to the control of the airplane. 

Two methods of driving remotely mounted pro- 
pellers were available: (1) drive-shafting with gears or 
(2) systems of chains or belts. The first would be 
inflexible, heavy, and costly, requiring new gears for 
each experimental change in the search for optimum 
engine/propeller speed ratio. The latter would permit 
a simple, inexpensive flexible range of speed ratios. 
Chain-drive systems were, at that time, frequently 
used in automobiles and other machinery and were 
readily available in a wide range of sizes. The Wrights 
had experience with chain-driven bicycles and chose 
to employ a chain drive with their propellers. Hence, 
a flywheel and chain sprockets were incorporated on 


Harvey H. Lippincott 

the engine for the propeller drive. Power loss in the 
transmission system was estimated to be 5 percent 
(Figure 1). 

propellers were similar in shape to those of 1903. 
Blades of IVz, 8, 8% and 10 inch widths were used. 
Rotational speed was increased with a sprocket ratio 
33: 10. Consequently, thrust was increased. A notation 
of 185 pounds at 400 revolutions per minute for the 
10-inch blade is mentioned in one notebook. 


1. Propulsion system installed in 1903 airplane with 
engine mounted on lower wing and chain drive to 
strut-mounted propellers. (SI neg. A38388) 

However, the chain drive was not without its 
problems. As Leonard S. Hobbs pointed out: 

Torque variations in the engine would tend to cause 
a whipping action in the chain, so that it was 
vulnerable to rough running caused by misfiring 
cylinders and, with the right timing and magnitude 
of normal regular variations, the action could result in 
destructive forces in the transmission system. This 
was the basic reason for the Wrights' great fear of 
"engine vibration," which confined them to the use 
of small cylinders and made a fairly heavy flywheel 
necessary on all their engines. 6 

The Wrights had progressed in their propeller work 
so that by June 1903 they had made their first flight 
propellers which were made of three l!/8-inch lami- 
nations of spruce, glued together. The propellers were 
shaped by hatchet and drawshave. Tips were covered 
with canvas and were 8V2 feet in diameter and 8 
inches in width, pitch 27 degrees, and had a blade 
area of 5.4 square feet. At an airplane speed of 24 
miles per hour and turning at 330 revolutions per 
minute the two propellers were designed to give a 
combined thrust of 90 pounds. Tip speed was 180 
feet/second. Efficiency was 66 percent. Sprocket ratio 
was 23:8 (Figure 2). Caldwell pointed out that 

the interesting thing about the Wright propeller, 
however, is that the speed of rotation was chosen in 
such a way as to give the most favorable conditions 
for propeller efficiency, due to the fact that the ratio 
of rotational tip speed to forward speed of the 
airplane was incorrectly chosen. 7 

In 1904 four different propellers were used on the 
second airplane including the 1903 propellers. All the 

2. Wright 1903 propeller used on the first flight. (SI 

Neg. 42363C and 42363A) 

The Wrights had found that propellers tended to 
twist under pressure and lose performance. In 1905 
they redesigned their propellers. To reduce the twist- 
ing, a pie-shaped portion of the leading edge was 
removed resulting in a relatively constant blade width 
for 30 percent of the distance from the tip to the 
center. These propellers were called "bent end" type. 
From now on the Wrights continued the use of "bent 
ends" with variations. Most propellers were 8V2 feet 
in diameter, though some were 9 feet in diameter. 
Blade widths ranged from 10 to 12 inches, pitch from 
23 degrees to 27 degrees at the tips. Thrust on the 
1905 "bent end" was 210 pounds at 450 revolutions 
per minute. Efficiency remained at 66 percent (Figure 

3. Wright 1905 "bent end" type propeller. (SI Neg. 

Propulsion Systems 


Construction of the "bent end" propellers was much 
the same as that of the 1903 type. The outer two- 
thirds of the blade was covered on both sides with 
light canvas which was glued on. The whole propeller 
was painted aluminum, lacquered, and polished. 

With the thrust/propeller system determined, an 
energy source to drive the propellers was required. 
For this, the Wrights chose the internal combustion 

The 1903 Engine 

The Wrights did not particularly want to get into the 
problem of designing and building an engine even 
though they previously had designed and built in 
1901 their highly successful shop engine. This was a 
single-cylinder stationary power plant producing some 
three horsepower, air cooled, and operated on city 
illuminating gas. 

With some hope that they could purchase a suitable 
engine, the Wrights sent a number of letters to various 
engine manufacturers inquiring for a gasoline engine 
developing eight to nine brake horsepower, weighing 
no more than 180 pounds, or an average of 20 pounds/ 
horsepower and free from vibration. While at least 
ten replies were received, none met the desired 

In view of the state of the art at the time, no 
suitable engines were available. An established man- 
ufacturer would hesitate to undertake the develop- 
ment of such an engine that would have little market. 
Furthermore, the cost and time to produce such an 
engine would be more than the Wrights could afford. 
Apparently, the Wrights pursued this avenue no 
further and continued their plans to design and build 
their own engine. 

To maintain cost control, the Wrights desired to 
machine engine parts as far as possible in their own 
shop. The metal-working machinery in the shop 
consisted only of a lathe, a drill press, and hand tools. 
Where complex machining operations were required, 
the Wrights had available in the Dayton area experi- 
enced machine shops with sophisticated equipment. 
Consequently, the meager equipment of the Wright 
shop added a design requirement to their engine. 
Work began upon the engine late December 1902. 

The Wrights assigned to their mechanic, Charles 
E. Taylor, the job of actually machining and con- 
structing the engine. He was more experienced in 
metal machining than they. Furthermore, their prin- 
cipal work was to be the construction of the first 
powered airframe. Little is known of the design 
process of the engine. According to Taylor, the 
Wrights included him in the many discussions and 
evaluations, but the decision making was theirs." 

Taylor said, "We didn't make any drawings. One 

of us would sketch out the part we were talking about 
on a piece of scratch paper and I'd spike the sketch 
over my bench. 9 No original drawings of the engine 
have ever been found. Furthermore, Orville Wright's 
diary of 1904 states, "Took old engine apart to get 
measurements for making new engine, an indication 
that drawings of the 1903 engine did not exist. 10 
Consequently, it would appear that Taylor, applying 
to the sketches his knowledge of machining and 
experience of fitting, was essentially responsible for 
establishing the detail dimensioning of the engine. 
Undoubtedly, much of the detail design of the engine 
was his. 

Studying the details of their engine, one is im- 
pressed in seeing the thought and consideration that 
went into the design. Obviously much forethought 
and careful design work preceded actual construction. 
There are no haphazard or "cobbled-up" solutions. 
All had been carefully laid out. For example, Taylor 

The first thing we did was to construct a sort of 
skeleton model, a test cylinder of about a four inch 
bore. The Wright boys were thorough that way. 
They wanted to see how some of the vital 
components worked before proceeding further. We 
hooked the test cylinder up to the shop power, 
smeared it with oil with a paint brush, and watched 
it run for short periods. It looked good; so we 
decided to go ahead with a four-cylinder model." 

This is all the more remarkable as the Wrights 
apparently made no detailed engineering drawings, 
only some temporary sketches. The intricate crank- 
case casting, for instance, shows evidence of much 
forethought on the overall engine design providing 
for webs, brackets, bosses, and contours to hold and 
support other engine parts. The records indicate that 
the Wrights made their own wood casting patterns. 

4. Wright 1903 engine chain-driven exhaust valve 
camshaft drive, valve boxes, and fuel line to air inlet 
can, spark advance/retard lever are visible. (SI Neg. 


Harvey H. Lippincott 





Drawings by 
J. H. Clark, A.R.Ae.S. 


I and 2. Bearing caps in one piece 
with plate 3. 

3. Plate screwed over hole 4 tn crank- 
case end. 

4. Key-shaped hole as hole S in inter- 
mediate ribs. 

6. Inter-bearings cap (white-metal 
lined) and screwed to inter-rib 
halves 7. 

8. Splash-drip feed to bearings. 

9. Return to pump from each com- 
partment of crankcasc base 
(" sump "> via gallery 10 and pipe 
lo pump 11 underneath jacket. 

12. Oil feed from pump via rubber 
lube 13. 

13. Drip feeds to cylinders and pistons. 

14. Gear drive to pump 

15.. Bin-end nuts, lock-strip, and shims. 

Gudgeon-pin lock. 

Piston-ring retainer pegs. 

Cylinder liner screwing into jacket. 

Open-ended " can " admits air. 

Fuel supply. 

(Hoi) side of water jacket makes 

surface carburetter. 

Sparking plug (comprising positive 

electrode 23 and spark-producing 

makc-and-break 24). 

Lever attached to lever ■ 26 via 

bearing 27 screwed into chamber 

neck 28. 

Levers with mainspring 2^ and 

imerspring 30. and rocked by 

" cam " 31. 

Cam with another alongside (for 

adjacent cylinder). 

Positive busbar feed to ill four 


Assembly retai Ding-rings. 

Sealing disc. 

Exhaust outlet ports. 

Camshaft right along on underside 

of jacket and also driving oil pump 



Spring-loaded sliding pinion drives 

make-and-break shaft 38 through 

peg in inclined slot 39 

Cam to push pinion 37 along and 

so alter its angular relation with 

shaft 38 (to vary timing). 

Exhaust-valve cams bear on rollers 

42 mounted in end of rocker-arms 


Generator floating coils. 

Friction-drive off flywheel. 

Sight-feed lubricator (on stationary 


Hardwood chain tensioner. 

5. Assembly cutaway of 1903 engine. (From Aeroplane, photocopied UTC Archive A-113) 

Probably these patterns served as three-dimensional 
drawings upon which design details could be worked 

In laying out the engine design, the Wrights wished 
to produce an engine that would provide the perform- 
ance necessary to fly their airplane. They were not 
interested in producing an engine of the highest 

technical refinement, as such would be both costly 
and time-consuming. Their first decision was to select 
the size, number, and arrangement of the cylinders. 
The decision was much influenced by a primary 
concern of engine roughness. Vibration from the 
engine would deleteriously affect the structural integ- 
rity of both a lightweight engine and a frail airframe. 

Propulsion Systems 


Their choice of four cylinders in-line was a practical 
choice, an arrangement serving well in the automotive 
field, and one that would provide reasonable smooth- 
ness in operation without undue complexity. It ap- 
pears that the Wrights considered vibration to come 
largely from explosion forces which would be miti- 
gated with multiple cylinders. They apparently did 
not take into account, or chose to ignore, the inherent 
shaking force in the four in-line arrangement which 
was about 91 pounds per cylinder (See Appendix 1). 
They chose a "square" four-by-four-inch size cylinder 
for reasons not recorded. However, this size provided 
a cylinder large enough not to penalize weight and 
small enough to ensure successful operation. This 
size, or slightly larger, was to remain standard on all 
subsequent Wright engines. The engine's displace- 
ment was 201 cubic inches (Figures 4 and 5). 

"The Mean Effective Pressure (MEP) in the cyl- 
inder, based on their indicated goal of 8 hp, would 
be a very modest 36 psi at the speed of 870 rpm at 
which they first tested the engine, and only 31 psi at 
the reasonably conservative speed of 1000 rpm." 12 
"Assuming a rich mixture, consumption of all the air, 
and an airbrake thermal efficiency of 24.50% for the 
original engine, the approximate volumetric efficiency 
of the cylinder is calculated to have been just under 
40%. " u (See Appendix 2). 

The matter of cooling was also an important decision 
at this time. Their shop engine was air cooled. Air 
cooling had many weight-saving considerations. Yet, 
the Wrights opted for water cooling. At that time 
water-cooled engines were predominant in automo- 
biles and were providing the best performance in 
general service. All subsequent Wright engines were 
of the water-cooled type. 

The decision to arrange the engine in a horizontal 
position rather than the usual vertical position probably 
was to reduce drag or to spread the engine weight 
over a larger mounting base, or both. 

On the detail design of the engine one of the most 
important decisions the Wrights had to make was that 
of the crankcase design and its cylinder arrangement. 
The Wrights were radical in choosing en-block con- 
struction. Most engines of that period had individual 
cylinders mounted upon the crankcase. In spite of 
the more complex design that an en-block, one-piece 
crankcase entailed, it eliminated the making and 
joining of a number of smaller pieces, hence saving 
both time and money. While machining of the crank- 
case may have required sending it to an outside 
machine shop, Taylor states that he bored out the 
cylinder mounting bores on the shop lathe. 14 Most of 
the machine work was of a nature which allowed it 
to be accomplished at the Wright shop. For lightness, 
the crankcase was cast in aluminium alloy of 8 percent 
copper and 92 percent aluminum. At that time alu- 

6. 1903 disassembled cylinder with barrel, valve box, 
valves in cages, and rocker arm. (SI Neg. 38626G) 

minum castings were used in many automobile engine 
crankcases. The Wright crankcase was designed so as 
to carry the major portion of the engine within, adding 
both strength and lightness to the engine. The open 
upper portion of the crankcase was covered with a 
screwed-on sheet steel plate. 

Cast iron was employed for the cylinder, which was 
machined as thin as practical to save weight (Figure 
6). The barrel was partially closed at the head end, 
terminating in a threaded boss. Installed through the 
interior of the crankcase, the barrel was screwed into 
the cylinder mounting bore seating on the outer and 
inner flange against the crankcase, thus forming the 
water jacket. A cast cylindrical valve box lay across 
the cylinder barrel, which contained the inlet and 
exhaust valves in cages. The valve box was screwed 
onto the threaded cylinder head boss, firmly tightening 
the barrel outer flange against the crankcase. The 
boss bore and the valve box formed the combustion 
chamber. Interestingly, the two cylinder barrel thread 
sets and flanges had to be accurately cut and fitted so 
that the screwing of the barrel into the crankcase and 
the valve box occurred in such a way to form a tight 
three-way joint. The valve box was strictly air cooled, 
which, however, became red hot after a few minutes' 
run (Figure 7). 

The inlet valve was an "automatic" or suction- 
operated valve, which eliminated one complete me- 
chanical valve-operating mechanism. The exhaust 
valve was mechanically operated by cams and rocker 
arms. The camshaft was driven by the crankshaft 
through a bicycle chain. Cast iron was employed for 
the entire valve system, except for cold-rolled steel 
valve stems. 

The power machinery — that is, crankshaft, con- 
necting rods, and pistons — was all made in the Wright 
shop. From a solid slab of heat-treated, relatively high 
carbon steel, Taylor cut and turned on the lathe the 


Harvey H. Lippincott 

7. Cross section of 1903 engine showing cylinder and 
valve assembly in crankcase. (Science Museum, 
London, from UTC Archive A-112) 

9. Underside of 1903 engine showing exhaust valve 
rocker arms, camshaft, spark advance/retard rod, and 
later added oil pump and lines. (SI Neg. 38626F) 

four-throw crankshaft. The shaft was an orthodox 
straight pin and cheek type without counterweights. 
Large plain babbitted bearings on each side of the 
cranks provided added stiffness. The Wrights were 
fortunate that the vibratory forces were relatively 
small in their engines so that the natural vibration 
frequency always fell outside the small operating 
speed range of the engine. A machined 26 pound cast 
iron flywheel, balanced against the magnitude of the 
explosion force, was shrunk onto the crankshaft (Fig- 
ure 8). 

The connecting rods were made from seamless steel 
tubing and screwed and pinned into phosphor-bronze 
cast ends drilled through to form the bearings. For 
the first engine, this rod construction was satisfactory 
but proved to be troublesome in later engines. In the 
six-cylinder engine it was replaced by a rod of I 

8. 1903 crankcase, open, showing power machinery. 

(SI Neg. 38626H) 

The pistons were typical of the time, machined 
from cast iron with long skirts and accommodation for 
three cast iron piston rings. 

The Wrights were very careful to control weight in 
the engine. Every part was weighed and recorded as 
it was finished so as to conform to their predetermined 
weight allowance. 

As first built, the 1903 engine was splash lubricated 
internally and externally, with liberal applications from 
an oil can. When the engine was rebuilt in 1916, 
another or new crankcase was obtained to replace the 
original, which had been broken after the first flight 
and apparently melted down to make other castings. 
In the replacement crankcases an oil pump and 
pressure oil system was incorporated, which sprayed 
oil over the thrust-loaded side of each piston (Figure 

The fuel system was relatively simple. There was 
no carburetor as such. Fuel flowed by gravity from a 
0.4 gallon gasoline tank mounted high on a strut 
through copper tubing to the engine. Two valves 
were placed in the fuel line; one was an on-off shut 
off valve, while the other was a regulating valve that 
could be adjusted to govern engine speed and power. 
Fuel was discharged into the top of a sheet metal 
"can," where it mixed with the incoming air and 
passed into an induction chamber. The mixture cir- 
culated over the hot surface of the crankcase, thereby 
vaporizing the mixture. The fuel/air mixture then 
passed into the intake manifold that surrounded the 
inlet valve and cage and on into the cylinders. 

Two ignition systems, the high-tension spark plug 
and the low tension make-and-break, were in general 
use in 1902. The high-tension system was somewhat 
simpler but less reliable than the more complex make- 
and-break. Probably for reasons of reliability and the 
fact that all the parts could be made in the Wright 
shop or easily procured, the Wrights chose the low- 

Propulsion Systems 


tension make-and-break system. Also, this system was 
not susceptible to fouling from excess oil in the 
cylinders, an important consideration for a new engine 
design. The spark was made by the opening and 
closing of two contact points inside the combustion 
chamber. An insulated, fixed contact point in the 
combustion chamber was connected to the Miller- 
Knoblock 10 volt 4 amp D.C. magneto which was 
friction-driven by the flywheel. A movable contact 
point inside the combustion chamber was mounted 
on an arm whose shaft passed through the cylinder 
wall. The shaft was attached to a steel spring leaf 
that rode over a cam on the ignition camshaft. The 
camshaft was driven by the exhaust valve camshaft. 
Through a mechanism, control of the spark timing 
was accomplished by advancing or retarding the ig- 
nition camshaft in relation to the exhaust valve cam- 
shaft. Dry batteries and a coil were connected for 
starting but were not part of the airplane. 

The cooling system consisted of a vertical radiator 
made from slightly flattened metal tubes connected 
to the engine water jacket with rubber "in" and "out" 
hoses. The cooling water circulated by convection. 

The engine was completed in February 1903 and 
first run on February 12. The Wrights were pleased 
with its operation and smoothness. Then began a 
period of break-in and development. In the course of 
development, stronger valve springs were installed 
which increased power to nearly 16 horsepower and 
reduced fuel consumption by half. Orville wrote: 

Due to the preheating of the air by the water 
jacket and the red-hot valves and boxes, the air was 
greatly expanded before entering the cylinders. As a 
result, in a few minutes' time, the power dropped to 
less than 75 percent of what it was on cranking the 
motor. The highest speed ever measured was 300 
turns (1,200 rpm) in the first fifteen seconds after 
starting the cold motor. The revolutions dropped 
rapidly and were down to 1,090 rpm after several 
minutes' run. 15 

According to the Wrights' measurements and cal- 
culations, the powers obtained at these speeds were 
15.76 horsepower and 11.81 horsepower, respec- 
tively. 16 

For their first engine, the Wrights set out to produce 
one of 8 horsepower not to weigh more than 180 
pounds. They calculated that this would be adequate 
to fly their machine. That they met this requirement 
is attested in a letter Wilbur wrote to George Spratt 
dated February 28, 1903: 

We recently built a four-cylinder gasoline engine 
with 4" stroke, to see how powerful it would be, and 
what it would weigh. At 670 rpm it developed H'A 
horsepower, brake test. By speeding it up to 1,000 
rpm we will easily get 11 horsepower and possibly a 
little more at still higher speed, though the increase 
is not in exact proportion to the increase in number 

of revolutions. The weight including the 30-pound 
flywheel is 140 lbs. 

Orville stated that the engine weighed 161 pounds 
dry and 179 pounds with the magneto. 17 "Complete 
with magneto, radiator, tank, water, fuel, tubing and 
accessories the powerplant weighed a little over 200 
pounds." 18 The Wrights met their objective, and 

With a suitable and satisfactory engine, the Wrights 
went on to achieve their first powered flight with it 
on December 17, 1903. 19 

It is remarkable that these two men, aided by 
Taylor, inexperienced in internal combustion engine 
design and especially multi-cylinder engines — much 
less in extra light construction — could, in two 
months, bring through an engine which was both 
operable and somewhat lighter than their 
specification. 20 

The 1904 Engines 

To continue flying and perfecting their mastery of 
flight, the Wrights in 1904 undertook to build a new 
airplane and engines to power it. Three more powerful 
engines were put into construction. Two engines were 
based upon the 1903 engine design, while the third 
was one of eight cylinders. Little is known about this 
eight-cylinder engine. As the four-cylinder engines 
were producing sufficiently more power, the eight- 
cylinder engine was not completed and the parts were 

The first 1904 four-cylinder engine was merely an 
improved 1903 and was designated No. 2. It had a 
bore of 4Vs inches, an oil pump, a fuel pump, and a 
10 percent increase of cylinder barrel water cooling. 
Its displacement was 214 cubic inches. Its one unusual 
new feature was the addition of a compression release 
mechanism. The pilot actuated a sliding shaft that 
forced stops under the exhaust valve rocker arm rollers 
and kept the valves open thus effectively shutting off 
the engine. Why the Wrights incorporated a compres- 
sion release has never been fully explained. It was, 
however, useful in starting and stopping. It was 
continued on all subsequent engines (Figure 10). 

Initially, the No. 2 engine developed 15 to 16 
horsepower. By 1905 it was developing 20 to 21 
horsepower. Most of the subsequent gain was due to 
the smoothing of the cylinders and pistons by wear. 21 
This engine was employed in the 1904 airplane and 
also in the 1905 airplane. 22 

Engine No. 3 was built in 1904 for experimental 
purposes. The two objectives, in providing this en- 
gine, were to improve performance and reliability. 
The engine was operated extensively on a test stand 
from 1904 until mid-1906. It had a four-inch bore and 
incorporated compression release and exhaust ports 
in the cylinder barrels. Several holes were drilled in 


Harvey H. Lippincott 

10. A 1904 No. 3 experimental engine with 
compression release mechanism. (UTC Archive A-108) 

the barrel which were uncovered by the piston at the 
end of its stroke. The purpose of the exhaust ports 
was to reduce the pressure and temperature of the 
exhaust gas passing through the exhaust valve. Ex- 
haust ports were also used on the subsequent four- 
cylinder vertical engines. By 1906, No. 3 engine was 
developing 24 to 25 horsepower at 1300 revolutions 
per minute, twice the power of the original 1903 
engine of the same size. 23 

The Four-Cylinder Vertical Engine 

In 1906, while still doing general development work 
on the flat experimental engine, the Wrights started 
two new engines, and for the first time the brothers 
engaged in separate efforts. One was "a modification 
of the old ones" by Wilbur and the other, "an entirely 
new pattern" by Orville. There is no record of any 
of the features of Wilbur's project or what was done 
in connection with it. 

11. Four-cylinder vertical engine (1906-12), right side 
showing fuel system injection pump and line to fuel- 
air manifold, and oil system pressure lines from sump 
to pump to crankcase and oil fill pipe. (SI Neg. 


There is a brief entry on it in Wilbur's diary but 
no further mention of it. 24 Orville's design was a four- 
cylinder vertical engine which was to become the 
most used of any model they produced. Production 
of this engine continued into 1912. From now on 
there are indications that Orville became the leader 
in engine design. 

Orville's engine represented significant change from 
the previous horizontal type. Hobbs concluded that 
"putting the engine in an upright instead of flat 
position, was probably done primarily to provide for 
a minimum variation in the location of the center of 
gravity with and without a passenger." 25 

Initially the four-cylinder vertical engines delivered 
28 horsepower at 1325 revolutions per minute and 
weighed 160 pounds dry. Throughout its production 
run many changes and improvements were made so 
that its power ultimately reached 40 horsepower at 
1500 revolutions per minute at a dry weight of 180 
pounds. These engines had a bore of 4% inches, a 
stroke of 4 inches, and a displacement of 240 inches. 
The engines generally were reliable though trouble 
was often experienced with exhaust valves, pistons, 
and cylinders (Figure 11). 

The most drastic change from the previous engines 
was the employment of individual cylinders and the 
abandonment of en-block construction (Figure 12). 
Again the sturdy aluminum alloy crankcase was cast 
in one piece with one side open and covered with a 
screwed-on sheet-steel plate. The cylinders were 
bolted to the flat top of the crankcase. However, 
many of the details were similar to those of the original 

12. Left side view of four-cylinder vertical engine 
with crankcase open to expose the power machinery. 
Compression release rod is mounted along top of 
crankcase. Mea magneto installed. Water circulation 
pump at far left. (P&W photo, UTC Archive D-15006) 

Propulsion Systems 


13. Four-throw crankshaft cut from steel blank and 
shrunk-on flywheel. (P&W photo, UTC Archive D- 

14. Tubular connecting rod screwed into cast ends. 
Pointed oil scupper is mounted on crankshaft bearing 
cap. (P&W photo, UTC Archive D-14992) 

of the connecting rod threw oil onto the cylinders. 

Gravity-fed gasoline from a tank mounted high in 
the struts went to a camshaft-driven fuel pump that 
metered the fuel, as there was no carburetor. The 
fuel entered a baffled inlet manifold where it mixed 
with air and was heated by the exhaust impinging on 
the manifold before entering the cylinder. 

A water circulation pump driven by the crankshaft 
was incorporated. Cooling water was piped to a 
horizontal manifold at the top of the cylinders and 
after circulation around the cylinders was collected in 
another horizontal manifold on the other side of the 
cylinders for return to the radiator. A normally supplied 
vertical tube radiator weighed 40 pounds and carried 
25 pounds of water. 

Now the Wrights installed a high tension ignition 
system. A Mea magneto, though sometimes a Bosch, 
was mounted on a bracket, cast integral on the 
crankcase and driven through gears from the camshaft. 
One spark plug was installed in each cylinder. By a 
foot pedal, the pilot controlled the spark advance 
which was the only speed control for the engine. 

A revised compression release mechanism was in- 
stalled. It lifted a collar on the pushrods to open the 
exhaust valve and stop the engine. 

The four-cylinder vertical engine was the only 
engine licensed by the Wrights for manufacture by 

engine, even though involving some modifications 
and improvements. Such parts were the crankshaft 
and flywheel (Figure 13), pistons, two-piece inlet and 
exhaust valves, rocker arms, and connecting rods 
(Figure 14). See Appendix IV for material composition 
and hardness of these parts. 

The cylinder was a complete one-piece iron casting, 
machined all over. Inlet and exhaust ports and valve 
guides were cut into an integral boss on the head of 
the cylinder. The inlet valve again was an "automatic" 
type, while the exhaust valve was mechanically op- 
erated. The exhaust rocker arm rocked on a steel stud 
screwed into the cylinder head and was actuated by 
a pushrod operated by the cam shaft inside the 
crankcase. A four-cornered flange near the bottom of 
the cylinder provided for fastening the cylinder to the 
crankcase. A cast aluminum water jacket was shrunk 
onto the cylinder barrels and covered about two-thirds 
of the barrel. The lower part of the barrel was not 
cooled because of the exhaust ports drilled into the 
barrel wall. Also, the cylinder had remained uncooled 
(Figures 15 and 16). 

A pressurized lubrication system was incorporated. 
An oil pump driven by the camshaft pumped oil from 
the sump and forced it through drilled passages in 
the crankshaft and crankcase to lubricate the bearings, 
cams, and cylinders. A scupper on the crankshaft end 


15. Cylinder assembly, push rod, rocker arm, and 
camshaft. (P&W photo, UTC Archive D-14996) 

ft --^ 

16. Disassembled cylinder showing cast iron barrel, 
cast aluminum water jacket, and two-piece valves. 
(P&W photo, UTC Archive D-15001) 


Harvey H . Lippincott 

17. Propulsion system of 1910 Baby Grand Racer 
with eight-cylinder engine installed. (SI Neg. # A- 


18. Early six-cylinder engine (1911) with fuel/air 
mixing manifold and exhaust ports adjacent to inlet 
parts. (U.S. Air Force Museum, phorocopied from UTC 
Archive A-lll) 

others. In Germany these engines were manufactured 
by Neue Automobil-Gesellschaft and in France by 
Bariquand et Marre. While no production figures exist, 
recollections of a factory foreman indicates a hundred 
engines were manufactured, if not more. 

The Eight-Cylinder Racing Engine 

The Wrights decided to enter an airplane in the 1910 
Gordon Bennett Cup Race. A small racing airplane 
was built. The 30 horsepower four-cylinder vertical 
engine was considered to be too small. They designed 
and built a 60 horsepower Vee-eight of 481 cubic inch 
displacement. It was essentially a combination of two 
standard four-cylinder engines on a special crankcase 
spread 90 degrees apart. The connecting rods were 
placed side-by-side on a modified four-cylinder crank- 
shaft. A single camshaft operated all the exhaust 
valves. The engine weighed about 300 pounds (Figure 

Unfortunately, the engine had a short life as the 
airplane was demolished before the race. Little in- 
formation on it survives. 

The Six-Cylinder Vertical Engines 

By 1911, as aircraft performance improved, the Wrights 
faced the necessity of equipping their airplanes with 
more powerful engines in the race to keep up with 
theircompetition. They had three choices: (1) increase 
the size of their four-cylinder engine; (2) develop 
their eight-cylinder engine; or (3) produce a six- 
cylinder engine. Dimensional increase in bore and 
stroke of the four-cylinder engine would increase 
operational roughness which the Wrights did not want. 
An eight-cylinder engine would probably be too large 
for their needs at that time. Their choice of a six- 
cylinder arrangement gave them the desired power 

19. 6-60 six-cylinder engine (1912-13) exhaust side. 

(SI Neg. 74828) 

with a very smooth operation. Also, it continued their 
experience with in-line engines. 

The cylinder bore remained at 4 3 /s inches. However, 
the stroke was lengthened Vi inch to AVz inches in 
order to increase displacement, which was 406 cubic 
inches. Early six-cylinder engines produced 50 horse- 
power (Figure 18). In 1913 the engine was redesigned. 
Called the 6-60, it gave 60 horsepower (Figure 19). 
The final version, the 6-70, produced 75 horsepower. 
One may ask why the Wrights did not produce more 
powerful engines, as did their competition who were 
turning out 80-100 horsepower engines. The answer 
probably lies in the high efficiency of their propellers, 
which required less power to produce the same thrust. 

While many of the four-cylinder engine parts were 
utilized in the six-cylinder engine, there were several 
significant differences and improvements. The cast 
aluminum crankcase was two pieces split along the 
horizontal crankshaft centerline. The upper crankcase 
was open on one side covered with a screwed-on 

Propulsion Systems 


sheet-steel plate. The crankshaft was supported in 
the upper crankcase only. The lower crankcase sup- 
ported the engine and provided for its mounting 
(Figure 20). 

20. Six-cylinder two-piece crankcase. (P&W photo, 
UTC Archive D-15015) 

A major improvement was in the cylinders. The 
three-piece cylinder consisted of a machined, cast 
iron barrel with a section of seamless steel tube shrunk 
on to form a water jacket, while a cast iron head was 
shrunk on to the top of the barrel. For the first time 
the head was water cooled. Four long studs running 
from the crankcase to the top of the cylinder head 
held the cylinders in place. Also, for the first time, 
one-piece forged steel valves were used, though 
strangely only for the inlet. The more critical exhaust 
valve was still the old two-piece type with a cast iron 
head and steel stem. Exhaust ports in the lower part 




21. Cylinder and valves for early six-cylinder engine. 

(P&W photo, UTC Archive D-15014) 

of the barrel were no longer used. On the early sixes, 
automatic inlet valves were employed. These were 
replaced with mechanically operated valves in the 6- 
60. The exhaust valves had the same compression 
release mechanism as the vertical four (Figure 21). 

The power machinery followed similar lines of the 
vertical four. The pistons were shortened an inch, 
saving 40 percent of their weight. For the first time 
forgings were used in the crankshaft and I-section 
connecting rods (Figures 22 and 23). 

22. Six-cylinder forged crankshaft and flywheel. 

(P&W photo, UTC Archive D-15018) 


23. Six-cylinder forged I section connecting rod and 
piston. (P&W photo, UTC Archive D-15017) 

Early engines continued use of the fuel injection 
pump and manifold. Vaporization problems apparently 
were encountered in the longer manifold which now 
was unheated. The old system was replaced in the 6- 
60 with two float-feed Zenith carburetors, each feeding 
three cylinders (Figure 24). 

For the 6-60 the cylinders were redesigned so that 
inlet and exhaust ports were on opposite sides of the 
cylinder, rather than side-by-side as on the earlier 
sixes. This arrangement facilitated the installation of 

The six-cylinder engine was a vastly improved 
engine, especially in its final form. Aesthetically, it 

Harvey H. Lippincott 

24. 6-60 six-cylinder engine with twin Zenith 
carburetors. (SI Neg. 74829) 

was an attractive engine showing spare design, sim- 
plicity, and utility. 

Concluding Remarks 

Summing up, Hobbs said: 

Overall, the Wright engines performed well, and in 
every case met or exceeded the existing 
requirements. Even though aircraft engines then 
were simpler than they became later and the design- 
development time much shorrer, their performance 
stands as remarkable. As a resulr, the Wrights never 
lacked for a suitable powerplant despite the rapid 
growth in airplane size and performance, and the 
continual demand for increased power and 
endurance. 26 

As with all machines, there were problems and 
failures. Lubrication was a continuing problem, es- 
pecially in the early years. Piston and cylinder barrel 
bearing surfaces had much distress with frequent 
scuffing and seizures, apparently due largely to poor 
lubrication and tight clearances. Broken and cracked 
cylinders were frequent. Valve failures were a constant 
annoyance. Part of the valve problem was caused by 
the cam design, which created rapid valve opening 
accelerations and high valve seating velocities. The 
Wright engine problems were no worse, and perhaps 
not as severe, as those of their competition. 

Nevertheless, the Wright brothers showed remark- 
able engineering talent in the field of propulsion. 
With no prior experience, they carefully analyzed the 
problem, as they had with their airplane, and devised 
good, basic engines of simple, sound design. Hobbs, 
one of the great aero engine engineers of all time, 
evaluated the Wrights' engineering. 

For the engineer particularly, the fascination of the 
Wrights' engine story lies in its delineation of the 
essentially perfect engineering achievement by the 
classic definition of engineering — to utilize the 

available art and science to accomplish the desired 
end with a minimum expenditure of time, energy 
and material. 27 

Harvey H. Lippincott, a native of Moorestown, New 
Jersey, holdsaB.S. in Mechanical Engineering from Georgia 
Institute of Technology. He held various positions in 
product support and marketing of aircraft engines with Pratt 
and Whitney Aircraft from 1941 to 1972 when be became 
Corporate Archivist for United Technologies. In 1959 he 
founded the Connecticut Aeronautical Historical Association 
and served as president of that organization for eight years. 
He has collaborated with many aviation museums and in 
1961 founded the Bradley Air Museum and served as its 
director. He is Chairman and member representing eastern 
United States of the Steering Committee for the Aviation 
Museum Directors and Curators Committee For the Inter- 
national Association of Transport Museums. He serves on 
the State Historical Records Advisory Board and the North- 
east Document Conservation Center Advisory Committee. 
He is a member of New England Archivists and Association 
for Study of Connecticut History. 


1. Frank W. Caldwell, "Notes on the History of Aircraft 
Propellers," circulated unpublished tract, Hamilton 
Standard Series, United Technologies Archives, May 1, 
1939, p. 1. Caldwell was Chief of the Propeller 
Engineering Division, Army Air Service, 1916-28; Chief 
Engineer, Standard Steel Propeller Company (later 
Hamilton Standard Propeller Company) 1929-35; 
Engineering Manager, Hamilton Standard Propeller 
Division, United Aircraft Corporation 1935^40; Director 
of Research, United Aircraft Corporation 1940-54; and 
winner of the Collier Trophy for 1933 for development of 
the controllable pitch propeller. 

2. Orville and Wilbur Wright, "The Wright Brothers' 
Aeroplane," The Century Magazine, September 1908, Vol. 
76, pp. 641-50. 

3. Orville Wright, "How We Made the Flight," Flying, 
December 1913, Vol. 2, pp. 10-12, 35-36. 

4. Ibid. 

5. Marvin W. McFarland, ed., The Papers of Wilbur and 
Orville Wright, McGraw-Hill Book Co., New York, 1935. 
Note bottom p. 617. 

6. Leonard S. Hobbs, "The Wright Brothers' Engines 
and Their Design," Smithsonian Annals of Flight, Number 
5, Smithsonian Institution Press, Washington, D.C. 1971, 
p. 12. Hobbs joined Pratt & Whitney Aircraft in 1927 as 
Research Engineer, advanced in the Engineering 
Department to become the Engineering Manager in 1935, 
and became Vice President for Engineering at United 
Aircraft Corporation 1944-56. 

7. Caldwell, p. 1. 

8. Fred F. Marshall, "Building the Original Wright 
Brothers Engine," Slipstream Aviation Monthly, May 1928, 
Dayton, Ohio. An article based upon a personal interview 
by the author with Charles E. Taylor. 

9. Charles E. Taylor as told to Robert S. Ball, "My Story 
of the Wright Brothers," Collier's, December 26, 1948, 
Vol. 122 (26), p. 27. 

10. McFarland, p. 414; Orville Wright's Diary, Thursday, 
January 14, 1904. 

11. Marshall. 

12. Hobbs, p. 10. 

Propulsion Systems 


13. Hobbs, p. 20. 

14. Taylor, p. 68. 

15. McFarland, p. 1210-13. 

16. Ibid. 

17. Orville Wright to Charles L. Lawrence, November 
15, 1928. 

18. McFarland, p. 1212. 

19. The restored 1903 engine is installed in the restored 
1903 airplane on display at the National Air and Space 
Museum, Smithsonian Institution, Washington, D.C. The 
engine was restored in 1916. In 1906 the crankshaft and 
flywheel were loaned to the Aero Club of America for 
their first aero show in New York. They were never 
returned. Therefore, the 1904 crankshaft and flywheel 
were installed in a new crankcase, which incorporated an 
oil pump and pressure oil system. 

20. Hobbs, p. 27. 

21. Orville Wright letter to Charles L. Lawrence, 
November 15, 1928. 

22. The 1904 No. 2 engine was restored in 1948 without 
a crankshaft and installed in the restored 1905 airplane on 
display at Wright Hall, Carillon Park, Dayton, Ohio. 

23. The No. 3 engine is on display at the Engineers 
Club in Dayton, Ohio. 

24. Hobbs, p. 34. 

25. Hobbs, p. 34. 

26. Hobbs, p. 59. 

27. Hobbs, p. 61. 

Appendix 1 

Balance of Wright 1903 Engine 

In 1966 L. Morgan Porter, Analytical Engineer, Pratt 
& Whitney Aircraft, analyzed the balance of the 
Wright 1903 engine for Leonard Hobbs's study of the 
Wright brothers' engines and reported: 

This four cylinder in-line engine of 4 inch bore and 4 
inch stroke had a five main bearing flat crankshaft 
with cranks 1 and 4 at 180 degrees from 2 and 3. 
With this type crankshaft the primary inertia forces 
and resulting rocking couples are inherently balanced 
among themselves. However, since this engine had 
no crankshaft counterweights, the primary inertia 
forces, in addition to the gas forces, are felt by the 
main bearings. At 1200 rpm this inertia force 
amounts to about 455 pounds in line with each 
cylinder, but with the direction at cylinders 1 and 4 
opposite to that at cylinders 2 and 3. 

For the second and all higher harmonics, there 
being no odd orders above the first, the inertia forces 
for all cylinders are in the same direction, and while 
this results in no rocking couples, there is a shaking 
force in line with the cylinders, the magnitude of 
which is four times that for one cylinder. At 1200 
rpm this secondary unbalanced shaking force 
amounts to about 91 pounds per cylinder, or a total 
force of 364 pounds. With the engine mounted 
horizontally in the airplane this resulted in a lateral 
shaking force on the engine supports occurring at 
twice crankshaft speed. Higher harmonics are usually 
neglected since their magnitudes become 
progressively smaller. The magnitudes of the primary 

and secondary inertia forces were determined in the 
following manner. 

Since actual reciprocating weights of the 1903 
engine were unavailable, an estimate was made from 
the following data for the 1908 engine 2 in which the 
bore had been increased to 4.375 inches. Piston and 
Rings, 4 lb. 7 oz.; Piston Pin, 8 oz.; Connecting 
Rod, Cap, Bolts and Nuts, 2 lb. 3 oz. of which 1 lb. 
was assumed to be reciprocating weight. 

Then for the 1903 engine, correcting the weight of 
the piston and rings directly as the cylinder bore, and 
taking the other weights the same (this neglects the 
increased connecting rod length of the 1903 engine, 
10 inches compared to 9.25 inches for the 1908 
engine), we have 

Piston and Rings (4.000/4.375)(4.4375) = 4.061b. 
Piston Pin plus Conn Rod Reciprocating 1.50 

Total Inertia Weight = 5.56 lb. 

The primary inertia force is given by 

Fj = 28.4 WiR( J CosG 


Where W, = Reciprocating Weight, lb 
R = Crank Radius, in 
N = rpm 
= Crank Angle From Top Dead Center 

The max. value is then 

Fj = 28.4 (5.56)(2)(1.2) 2 = 454.5 lb 

The secondary inertia force is given by 

F ' = 28 - 4W ' R (TSo)"(E |Cos2e 

Where L = Conn Rod Length, in 
The max. value is then 

F, = 28.4(5.56)(2)(1.2)M — J =90.7 lb 


1. Internal memorandum L. M. Porter to G. N. Cole, 
Chief Design Engineer, Pratt & Whitney, dated May 26, 
1966; Leonard S. Hobbs papers, Series 1, Smithsonian 
Monograph, Folder 22, United Technologies Archives. 

2. Porter had available a 4-cylinder vertical engine of 
1911 upon which an engineering analysis had been 

Appendix 2 

Volumetric Efficiency of Wright 1903 Engine 

In 1966 L. Morgan Porter, Analytical Engineer of 
Pratt & Whitney Aircraft, conducted a study of the 
volumetric efficiency of the 1903 engine by extrapo- 
lating data from tests conducted on a single cylinder 


Harvey H. Lippincott 

engine with a Pratt & Whitney R-2800 cylinder. Based 
upon Porter's study, 1 the volumetric efficiency ranged 
from about 37% to 42% and was derived as follows: 

Volumetric efficiency is defined as the ratio of the 
weight of air actually taken into the cylinder to the 
weight of air that would exactly fill the piston dis- 
placement at the inlet density. It is thus a figure of 
merit that can be used to compare the pumping ability 
of engines of different sizes or speeds. Volumetric 
efficiency is dimensionless, and may be greater or less 
than unity. 2 


NV dPl 

where M t = mass of fresh mixture per unit time 
N = number of revolutions per unit time 
V d = total displacement volume of the engine 
p = inlet density 

The denominator of this expression can readily be 
determined once the engine speed, displacement, 
and inlet air density are determined. For unsuper- 
charged engines the inlet density may be taken at 
standard conditions of 60°F and a pressure of 14.7 psi 
if the volumetric efficiency is to be based on conditions 
at the entrance to the carburetor. This will give the 
volumetric efficiency as a measure of the pumping 
ability of the entire engine. 

As for the numerator of the expression for volumetric 
efficiency, the only accurate way to determine this is 
to actually measure the weight of air consumed by 
the engine in a given time with all other conditions, 
such as fuel-air ratio, speed, power, etc., remaining 
constant. For the Wright engine with its ported 
cylinders the actual air measurement would be a 
problem. Furthermore, the power did not remain 
constant, falling from 16 to 12 hp after warm-up. 
Therefore, any value of volumetric efficiency for this 
engine based on brake horsepower and speed must 
be considered an approximation only. 

Porter arrived at a method that should give a fair 
indication of the volumetric efficiency of the 1903 
engine based on an assumed value for the brake 
thermal efficiency. 

Employing the equation: 

P = JM a (FQ C ) -n 

where P = power developed 

J = mechanical equivalent of heat 
M a = mass flow of dry air per unit time, or 

air capacity 
Q c = heat of combustion per unit mass of fuel 
r\ — thermal efficiency, which may be indi- 
cated or brake, depending upon whether 
P is defined as indicated or brake power 

F = fuel-air ratio 
and transposing the equation: 

M a 

J(FQ c h 

If the value of P is taken as 15.76 BHP as calculated 
by Orville Wright to be the maximum power ob- 
tained, 3 and Ricardo's FQ C value of 1290 BTU per 
pound as the heating value of air when burned with 
gasoline (based on a stochiometric mixture with fuel 
of about 19,000 BTU per pound). 


will give the weight of air consumed equivalent 

to the BHP output. Then dividing this by J of 778 
ft-lbs per BTU and by the brake thermal efficiency 
r\, an indication of the total air consumption of the 
engine is obtained. 

This method, of course, should only be used where 
the correct or rich mixtures are used, and depending 
on the accuracy of the air brake thermal efficiency (a 
function of the brake specific air consumption), will 
give the correct value for the air consumption for the 
BHP developed based on the assumption that all the 
air combines with the fuel. The air brake thermal 
efficiency (thermal efficiency based on air consump- 
tion) is used since any excess fuel is wasted and 
efficiency values based on the fuel consumption would 
be too low. 

With an assumed air brake thermal efficiency of 
about 24.5%, explained later, the total air consumption 

(15.76 BHP) (33,000 ft. lbs/min/HP) 

M = 

(778 ft-lbs/BTU)(1290 BTU/lbs air) (.245) 
lb. air 

M a = 2.115 


The volumetric efficiency for a four-stroke cycle 
engine, such as the 1903 engine, is expressed by the 

e v = 

NV dP , 

where Mi = mass of fresh mixture per unit time 
N = number of revolutions per unit time 
V d = total displacement volume of the engine 
pi = inlet density 
2 = two crank revolutions per cycle 

in which the mass of fresh mixture which passes into 
the cylinder in one suction stroke is divided by the 
mass of this mixture which would fill the piston 
displacement at inlet density. 

The engine displacement (4 inch bore, 4 inch 
stroke, 4 cylinders) is 201 cubic inches. Engine speed 

Propulsion Systems 


at maximum power is 1200 rpm; then the swept 
volume of the engine is: 

V H N 

201 in 3 1200rev/min 69.792 ft 3 


2 1728 in 3 /ft 3 2 rev mm 

Standard air density at the inlet is: 
p = 0.0764 lb. air/ft 3 
and when multiplied by the swept volume is 
'69.792 ft 3 \ /.0764 lbs. air\ 5.332 lbs air 


ft 3 


Then Volumetric Efficiency = 2.115/5.332 

= 0.397 or 39.7% 

Determination of the assumed air brake thermal 
efficiency (r\) was based upon data from experimental 
tests conducted on a single cylinder engine using a 
Pratt & Whitney R-2800 cylinder. 

To use the air brake thermal efficiency of the R- 
2800 single cylinder engine to estimate the air brake 
thermal efficiency for the 1903 engine, it would appear 
that this should be done at the 1200 rpm speed of 
the 1903 engine. However, 1600 rpm was the lowest 
speed at which the R-2800 single cylinder would 
operate satisfactorily at atmospheric conditions and a 
fuel-air ratio of .075. 

The assumed thermal efficiency of 24.5% for the 
1903 engine was arrived at by correcting the thermal 
efficiency of 28.95% at 1600 rpm for the R-2800 single 

cylinder, thought to be about optimum for atmospheric 
conditions, directly as the ratio of the air cycle 
efficiency at an assumed expansion ratio of 4.5 for the 
1903 engine to that for an expansion ratio of 6.75 for 
the R-2800. This gives (.452/.534)(.2895) = .2450 or 
24.5% brake thermal efficiency. 

Again assuming an expansion ratio of 4.00 for the 
1903 engine instead of 4.50 as used above, similar 
calculations lead to a brake thermal efficiency of 
23.09% and a volumetric efficiency of 42.1%. For an 
expansion ratio of 5.00 the results are 25.7% and 
37.7%, respectively. 

This analysis puts the volumetric efficiency of the 
1903 engine in the range of about 37% to 42%. The 
validity of these results depends, of course, on the 
assumption of a rich mixture, that all the air is 
consumed, and the accuracy of the assumed thermal 
efficiency. However, the results are believed to be 
the best available with the data at hand. 


1. Internal memoranda L. M. Porter to L. S. Hobbs and 
G. N. Cole, Pratt & Whitney Aircraft, dated July 15, 
1966, and August 19, 1966, Leonard S. Hobbs papers, 
Series 1, Smithsonian Monograph, Folder 22, United 
Technologies Archives. 

2. C. F. Taylor and E. S. Taylor, "The Internal 
Combustion Engine," International Textbook Co., 
Scranton, Penn. 1949. 

3. Marvin W. McFarland, ed., The Papers of Wilbur and 
Orville Wright, McGraw-Hill Book Co., New York, 1953, 
pp. 1210-13. 


Harvey H. Lippincott 

Appendix 3 

In 1966 Pratt & Whitney Aircraft disassembled and 
analyzed a Wright brothers 1911 four-cylinder vertical 
engine belonging to the National Air and Space 
Museum, Catalog No. M-1952-108. Material com- 
position and hardness tests were conducted on several 
major steel components. Results of these tests are 
given in the following table and are characteristic of 
Wright engines: 

Material and Hardness — Wright Brothers Engine 1908-1910 

Part Name 


Piston Pin 
Push Rod 
Rocker Roller 
Main Bearing 

Material — Approximate Composition 

1.5Cr 2.Ni .3Si .2 Mn Alloy Steel 
Carbon Steel 

1.5 Cr 3.Ni .3 Si .2 Mn Alloy Steel 
Carbon Steel 

6-8 Cu 6-8 Sb. .2Pb Bal. Sn 

as taken 








R-15T R-C 










Shaft Portion 
Unworn Cam Area 
Worked Cam Area 

Solid Babitt Shell 

Propulsion Systems 


Structural Design 

of the 1903 Wright Flyer 



When Wilbur and Orville Wright turned from making 
bicycles to making airplanes virtually only the powered 
flying machine remained an unsolved — and some said 
insoluble — problem. But the Wrights were convinced 
rational design was the key to proper construction of 
a flying machine. What sets their work apart from 
that of other would-be flying machine builders is their 
reliance on gliding experience and proven theory for 
design guidance. 

The Wright brothers' understanding of aerodynam- 
ics, stability, control, and propulsion is discussed at 
some length in other papers in this publication. This 
work is limited to their appreciation of the interplay 
among these topics during design and the way in 
which theory influenced the Flyer's structural form. 

From a design perspective, the structural form of 
the Flyer is not incidental. It is a consequence of the 
Wright brothers' understanding of the principles of 
air flow at the time of construction, of their choice, 
sizing, and disposition of materials in order to produce 
a machine strong enough yet light enough to fly with 
an engine of reasonable power, and of their great 
concern for safety. When all of these factors are taken 
into account we begin to appreciate how well the 
Wright brothers understood the art of aircraft design. 


By the latter part of the nineteenth century, structural 
analysis had emerged as one of the most advanced of 
the engineering arts. Methods of analysis common to 
all structural work were conveniently summarized in 

engineering handbooks such as Trautwein's Civil 
Engineers' Pocketbook of 1888 and Kents' Mechanical 
Engineers' Handbook of 1896. However, it remained 
to apply these general methods to design of a flight 
structure, and this requires a configuration and a 
means of estimating loads. 

The Wrights' preference for a biplane arrangement 
can be traced to Wilbur's concept for achieving lateral 
control by warping the lifting surface of a "double- 
deck" machine similar to that tested by Chanute and 
Herring in 1869-97 (Reference 1). When their biplane 
kite experiment of 1899 showed the concept could 
be made to work without loss of structural stiffness, 
the Wrights chose to use a biplane configuration for 
their manned gliders. Wilbur later elaborated on the 
choice in an address presented before the Western 
Society of Engineers in 1901. While discussing the 
problem of control, he cited the prior work of Chanute 
and stated, "The double-deck machine built and tried 
at the same time (by Chanute) marked a very great 
structural advance as it was the first in which the 
principles of the modern truss bridges were fully 
applied to flving machine construction" (Reference 

However, as used by Chanute the fully trussed 
biplane structure was incompatible with the Wrights' 
scheme for obtaining lateral control by warping the 
wings to present the right and left sides at different 
angles of attack. The Wrights' ability to modify the 
system of trussing to accommodate wing warp without 
compromising the structural integrity of the biplane 
cellule is a convincing example of their design inge- 


1. Three-view drawing of the Flyer 

nuity and familiarity with structural behavior. 

As finally evolved on the 1903 Flyer, shown in 
Figure 1, the central bays of the biplane cellule are 
completely trussed on all four sides to form a rigid, 
warp-free box, while the outer bays are rigidly trussed 
in the plane of the front spars only. Control wires, 
which also function as flying and landing wires, are 
used to truss the outer bays in the plane of the rear 
spars. With this arrangement the wing structure forms 
a complete load-carrying system for resisting the 
vertical loads and bending moments encountered in 
flight (Reference 3). 

The Wrights' method for resisting lateral and lon- 
gitudinal forces on the wing deserves particular atten- 
tion since their 1903 machine did not have a drag 
truss in the plane of the lifting surfaces. Instead, it 
used the fabric covering as a shear web. By orienting 
the warp and woof of the fabric at 45 degrees to the 
spars, the rectangular wing structure is made suffi- 
ciently stable to withstand lateral and longitudinal 
forces while retaining its abilitv to twist (Reference 

To enable the aft part of the wings to be warped 
to either an up or down position, the wing ribs were 
made in two parts with the plane of separation at the 
rear spar. The ribs were then rejoined with two thin 
flat strips of spring steel attached to the rib caps above 
and below the rear spar. These spring steel "hinges" 
improved reliability while reducing the amount of 
force required to warp the wings. 

A final touch of design detail is found in the hinged 
connectors that join the struts to the wing spars. By 
use of hinged connectors that function like universal 
joints the wing could be warped without imposing 
eccentric loads on the structural members. 

When all of these modifications to the Chanute 
trussed biplane structure are taken into account, one 
begins to better appreciate how closely Wilbur and 
Orville Wright paid attention to design detail. 

Undoubtedly the most distinctive (and controver- 
sial) feature of the Flyer is its forward mounted 
elevator. But the Wrights' choice of a canard config- 

uration is clearly explained in Orville Wrights' letter 
to Alexander Klemin on April 24, 1924 (Reference 
5). In Orville's words: "We originally put the elevators 
in front at a negative angle to provide a system of 
inherent stability. . . . We found it produced inher- 
ent instability. We then tried using our 1900 glider 
backwards with the rear edges foremost and found 
the stability much improved; but we retained the 
elevator in front for many years because it absolutely 
prevented a nose dive such as that in which Lilienthal 
and many others since have met their deaths." Thus, 
the Wrights' choice of a canard configuration is seen 
to have resulted from their concern for stall recovery 
and was a conscious decision in which safety played 
a dominant role. To them, control — not stability — 
was the main problem requiring attention. 

Preliminary Design 

Historians have always felt somewhat uncertain of 
how the Wrights went about designing their flying 
machines. A careful reading of The Papers of Wilbur 
and Orville Wright, edited by Marvin McFarland, 
reveals they started with a clearly defined objective 
and with this objective in mind prepared appropriate 
design specifications. Thus, Wilbur and Orville Wright 
functioned in much the same way as any project 
engineer today. 

On December 11, 1902 (Reference 6) Wilbur dis- 
closed their objective to Octave Chanute in a letter 
in which he writes: "It is our intention next year to 
build a machine much larger and about twice as heavy 
as our present machine (the 1902 glider). With it we 
will work out problems relating to starting and han- 
dling heavy weight machines, and if we find it under 
satisfactory control in flight, we will proceed to mount 
a motor." Apparently this objective was later modified 
for there is no record of any attempt to fly the 1903 
machine as a glider. However, Wilbur's statement 
suggests their plan was defined in order to enable 
them to obtain maximum benefit from their previous 
design experience with gliders. 

Their design specifications, shown in Figure 2, 
were extracted from Orville's letters to George Spratt 
on June 7, 1903 (Reference 7), and Charles Taylor, 
their mechanic, on November 23 (Reference 8) of the 
same year. These specifications indicate their original 
plan was to build a machine with a wing area of 500 
square feet. From their experience with gliders, it 
was estimated that a machine of this size would weigh 
625 pounds when fitted with a 200-pound engine. It 
was further estimated that the planned machine would 
be able to fly 23 miles per hour with an 8 hoursepower 
engine driving propellers able to produce 90 pound 
thrust when rotating at 330 revolutions per minute. 
It is only logical to ask how the Wrights arrived at 


Howard S. Wolko 




2. Preliminary goals 

500 sq.ft. 
62 5 lbs. 
8 hp. 

90 I bs. 

23 m ph. 

these particular estimates, and what the choice tells 
us about their understanding of aircraft design. 

The weight estimate of 625 pounds appears to have 
been strongly influenced by their experience with 
gliders and was determined as shown in Figure 3. 
Here the airframe weight of 425 pounds includes the 
weight of an operator (Wilbur weighed 140 pounds, 
Orville 145 pounds) but does not include an allowance 
for the weight of an engine. The wing loading of 0.85 
pound/foot 2 is for the manned airframe only and closely 
agrees with the wing loading of 0.83 pound/foot 2 used 
in design of their successful 1902 glider. The esti- 
mated engine weight of 200 pounds is from Orville's 
June 7 letter to George Spratt. When this engine 
weight is added to the weight of the manned airframe, 
the wing loading for the planned machine becomes 
1.25 pounds/feet. 2 

Once a reasonable weight estimate has been deter- 
mined, velocity can be estimated by considering the 
machine to be in steady flight in calm air. Under such 
idealized conditions of equilibrium, lift equals weight 
and velocity can be calculated from the expression 
for lift. In the form used by the Wrights during design 
of their 1903 machine, these idealized conditions were 



500 X .85 = 425 


TOTAL WEIGHT = 625 lbs 

1. with operator 

3. Weight estimate 

L = W = .0033 Sc,V 2 

where V = velocity in miles per hour 

S = total area of the lifting surface in ft 2 
c, = lift coefficient as determined from the 
Wrights' wind tunnel measurements on 
airfoil number 12. 
.0033 = the Wrights' "air presssure coefficient" 
(Reference 9) 

In designing their 1900 and 1901 gliders, the 
Wrights had estimated performance by expressing 
equilibrium in terms of the lift equation used by 
Lilienthal, Chanute, and others, that is, 

L = W = 0.005 Sc,V 2 

This expression contains the so-called Smeaton 
coefficient of 0.005 introduced by John Smeaton in 
1752 and gives an optimistically high value for lift. 
As it turns out, the Smeaton coefficient was wrong, 
but it had gone unchallenged and uncorrected for 
some 150 years. 

When the performance of the Wrights' 1900 and 
1901 gliders did not live up to expectations, they at 
first questioned Lilienthal's data on which their design 
was based. But when wind tunnel tests showed 
Lilienthal's data to be essentially free from error they 
concluded the fault was in Smeaton's coefficient. To 
bring their performance predictions into closer agree- 
ment with their glider experiments, the Wrights 
substituted data measured during flights with the 1901 
glider into the lift equation and calculated the more 
nearly correct coefficient of 0.0033. l 

Since the lift coefficient, q, was known to vary with 
the angle of attack, the Wrights' velocity estimate can 
be determined by choosing an angle of attack range 
of from 2.5 to 7.5 degrees, which corresponds to the 
Wrights' gliding experience, and constructing the 
table shown in Figure 4. Using the Wrights' wind 
tunnel results for surface number 12, the minimum 
velocity in this angle of attack range is seen to be 23 
miles per hour. This velocity represents a worst case 
condition corresponding to the velocity that had to be 
obtained for the machine to fly in the angle of attack 
range the Wrights' considered desirable. 

For steady flight in calm air, equilibrium also 
requires thrust to equal total drag. The drag equation 
used by Lilienthal, Chanute, and others is written 

D = 0.005 Sc d V 2 

The form of this expression is identical to that of the 
lift equation but the drag coefficient, c b is used in 
place of the lift coefficient, c,. Moreover, total drag 
is the sum of the drag attributed to the lifting surface 
and that due to the frontal area of the machine. In 
the form used by the Wright brothers, these equations 
are written 

Structural Design 



.0033 Sc, ^ V 2 


D f = .0033 SfC,- V 2 

T = D t = D + D f 

The only new terms introduced are S f , denoting 

the frontal area of the machine, which the Wrights 

c d 
estimated to be 20 ft 2 , and the drag/lift ratio — . Wilbur 

c i 

and Orville Wright used this somewhat unconven- 
tional form of the drag equation simply because they 
did not measure drag directly in their wind tunnel 
measurements. Instead, they measured the drag lift 

ratio, — (Reference 10). Constructing the table shown 

in Figure 5 shows the maximum value of thrust in 
the preferred angle of attack range to be 90 pounds. 
This also corresponds to a worst case condition for 
the angle of attack range of interest. 



L = W = ,oo33 Sc V 

V = 



oo33 Sc 

/ 625 


C l 


.14 5 




3 5.0 



2 7.2 



2 3.2 




4. Velocity estimate 

The Wright brothers' estimate of the power required 
is obtained by multiplying the total drag by the 
volocity. As shown in Figure 6, the maximum power 
required to fly in the preferred angle of attack range 
is 8 horsepower. 

It is indicative of the Wrights' consistent exercise 
of sound engineering judgment that they chose to 
base design of their first powered machine on worst 
case conditions. Had they not done so, it is quite 


D = .oo33 S c (_li) V 

D, = .oo33 S, c, (_S)V 

D t = D + D ( 


C | 


c l 




T=D t 

.14 5 

.2 6 3 

5 1.3 

1 66 


1 73 



.13 8 

3 5.0 





.5 15 

.10 5 

2 7.2 





.70 6 

,10 8 

2 3.2 






.11 8 





5. Thrust estimate 

likely the weight growth experienced during construc- 
tion of the 1903 machine would have rendered it 
incapable of flight. 


Many writers have claimed that the structural design 
of early aircraft had to be largely a matter of chance 
simply because no one, including the Wrights, had 
any real knowledge of flight loads. Supporters of this 
thesis contend analysis, if any, was limited. They 
further suggest that members were sized by the 
haphazard procedure of choosing the lightest possible 
member, judging proportions by eye, and then, by 
trial and error, replacing failed components with 
stronger members until satisfactory structural perform- 
ance was obtained. While some early experimenters 
indeed did resort to trial and error procedures, it 


P r = 

D t V 


D t 


P r 

mile lbs 



17 2.6 



2 3.6 



3 5.0 





2 7.2 

1 869 




2 3.2 

1 654 







6. Power estimate 


Howard S. Wolko 

would be a grave error to assume the Wright brothers 
were so inclined (Reference 11). 

In contrast with other experimenters, Wilbur and 
Orville Wright do not appear to have worried much 
about loads. To them, a load factor (erroneously 
thought to be a factor of safety) applied to external 
forces in equilibrium with the weight offered a com- 
mon sense solution to the problem. Guesses, backed 
by reasoning, still had to be made in order to account 
for the span wise distribution of lift. But once lift was 
reasonably apportioned the loads on individual com- 
ponents could be determined readily. It is quite likely 
the Wrights simply assumed lift to be uniformly 
distributed along the span. While this assumption is 
in error, it satisfied the load conditions required by a 
number of the methods of structural analysis subse- 
quently needed. 

Moreover, it simplified the task of estimating how 
much of the load was carried by the front and rear 
spars. In the wing cross section shown in Figure 7. 
W r denotes the running load which for a uniform 
distribution of lift is nothing more than the total load 
on the wing divided by the span. Wilbur and Orville 
Wright were aware of biplane effect from their wind 
tunnel tests, but it is doubtful that they knew how 
much load was carried by each wing. Consequently, 
it is quite probable they simply reduced the load 
carried by the lower wing. For the purposes of this 


R., = .51 w r 


*R 2 = .49 w r 

J W ' 

c p. at .5 c 

R,= ,19w r 
7. Spar loading 

R 2 =.81 w r 

work, the lower wing is considered to carry 85 percent 
of the load on the upper wing. Thus, the upper wing 
supports 340 pounds of the estimated total weight of 
625 pounds. This gives a running load of 8.8 pounds/ 
foot. By considering the running load concentrated at 
the center of pressure, the load carried by each spar 
can be determined readily. 

Since the center of pressure changes with angle of 
attack, two cases corresponding to extremes of flight 
in the preferred angle of attack range are shown. In 
the first case, the center of pressure is located at the 
30 percent chord line. This case corresponds to an 
angle of attack of 10 degrees. In the second case, 
which shows the center of pressure at the 50 percent 
chord line, the angle of attack is 2 degrees. For the 
worst case condition, in which the center of pressure 
is located at the 50 percent chord line, the rear spar 
of the upper wing is seen to carry 81 percent of the 

Sizing of Wing Components 

As mentioned earlier, engineering handbooks pub- 
lished in the latter part of the nineteenth century 
provided a convenient source for methods of analysis 
needed to size structural members. Although these 
methods were not always based on rigorous mathe- 
matical analysis, they were commonly used by nine- 
teenth-century engineers and did result in answers of 
acceptable engineering accuracy. For example, the 
method shown in Figure 8 was recommended for use 
by bridge builders confronted with the problem of 
finding the vertical shear at the supports of a contin- 
uous beam uniformly loaded along its length. By 
multiplying the given coefficients by the running load 
on the beam and the distance between supports, the 
vertical shear at each support could be determined 
with surprising accuracy. Moreover, the method could 
be extended to include any number of supports in 


No. Spans 

I 1 2 

Is. -_5.U -.2.1 

8 8 8 8 

I =1 3 

1-4. -_§_ l_5_ -il_S_ -^_l 

10 10 10 10 10 10 

I I 4 

I 11 — "I 7 I 15 — 1 3 I 1 3 -15M7 -11 I 

28 28 28 28 28 28 28 28 

I I 5 

| l5 -23J2O —18 | 19 — 19ll 8 — 2QI23 -15 I 

38 3838 38 38 3838 3838 38 

I _J - -6 

Ul -63_|s5 -49 | S 1 -53 |53 -51 [49 -ssl 63 -41 | 

104 104104 104104 104104 104104 104104 104 

8. Vertical shear coefficients 

Structural Design 


the following manner: For a beam with an odd number 
of spans, like five, follow down a given line of 
coefficients, such as the line on the right hand end, 
and the coefficient is seen to be i5 As. The 15 is 
obtained by adding the 1 1 of n /28 to the 4 of 4 /io. The 
38 is obtained by adding the 28 of u /zn to the 10 of 
4 /io. For a beam with an even number of supports, 
such as four, the end coefficient is seen to be u /z8. 
Here the 11 is obtained by multiplying the 4 of 4 /\o 
by 2 and adding the 3 of Vs. The 28 is obtained by 
multiplying the 10 of 1/io by 2 and adding the 8 of Vs. 

The origins of such questionable methods are a 
constant source of amazement but for a beam of eight 
spans, the results are within about five percent of 
those obtained by more exact methods. What is of 
importance to this work is that the problem is identical 
to that of finding the vertical shears on the spar of a 
biplane wing supported by interplane struts. Whether 
the Wrights resorted to the use of this empirical 
method or more exact, but mathematically cumber- 
some, methods is unknown. However, in passing it 
should be noted that this empirical method was still 
being used for aircraft design in Great Britain as late 
as 1917 (Reference 12). 

Once the vertical shears on the upper and lower 
spars of a wing truss were known, the forces in all 
truss members could be obtained with the aid of a 
widely used graphical procedure known as Maxwell's 
Diagram. For the purpose of this work, a Maxwell's 
Diagram was used in conjunction with the vertical 
shears obtained by using the empirical method of 
Figure 8 to determine the force in each member of 
half the rear truss of the Wright machine. The external 
loads on the half truss, shown in the upper diagram 
of Figure 9, correspond to the equilibrium condition 
of steady flight in calm air at a two degree angle of 
attack. These loads were used to construct the force 
diagram shown in the lower part of Figure 9. The 
force in each component is indicated on the upper 
diagram with a letter suffix to denote whether the 
member is in tension or compression. 

While construction of the Maxwell Diagram is 
straight forward, the tenets of truss theory preclude 
the method from providing any information on bend- 
ing moments. These must be obtained by using an 
independent procedure, which, for a continuous beam 
with eight bays, would casually imply use of the 
Method of Three Moments. However, the Method 
of Three Moments is mathematically cumbersome 
and susceptible to arithmetic error when the calcula- 
tions must be performed by hand. Consequently, an 
empirical method similar to that used for determina- 
tion of vertical shear was often used by nineteenth- 
century engineers. As shown in Figure 10, the bending 
moment coefficients must be multiplied by minus 
W r l 2 , that is, the negative product of the running load 


— i 

SCALE lbs 

9. Determination of component loads 


, No. Spans 


I I 2 

I I I 

1 -L o 

10 10 

I 3 

I I I 

3 2 3 

28 28 28 ° 

~1 4 

"3"8 7B 

38 38 

I I 

o -ii. 8 -5- -5- -LL 

104 104 104 104 104 ° 

!' I--6 

10. Bending moment coefficients 

and the square of the support spacing, to obtain the 
bending moments. As in the case of vertical shears, 
the method can be extended to include any number 
of supports using the same arithmetic procedure 
previously described. 

When used to determine the bending moments on 
the rear truss of the Wright machine, the procedure 
resulted in the bending moment diagram shown in 


NowardS. Wolko 


11. Bending moment diagram 


Column Load P^ r = — ^' 

Material Spruce 

E = 1.6 x 10° psi 


c r 

I = 

b = 

5 P = 1800 lbs 

b IV 

12 I 


TT 2 E 

= .205 in 

= 1.5 in 

b = 1.68 in 

Margin = 5.6 
12. Determination of spar width 

Figure 11. An independent check of the bending 
moment diagram using the Method of Three Moments 
resulted in the curve plotted with a broken line and 
indicates the maximum moment of —389 in-lb is well 
within engineering accuracy. The letters A through 
D designate locations of the interplane struts. 
In a biplane, the upper wing spars function as beams 
subjected to the simultaneous action of bending and 
compressive loads. Today's engineers are taught to 
be wary of such load combinations for, as the beam 

deflects due to bending loads, the compressive end 
load tends to interact, or couple, with the deflection 
so as to further increase bending. Although this 
behavior was understood by the late nineteenth cen- 
tury, the tendency was to regard it as a secondary 
effect. The cause of a number of early monoplane 
disasters can be traced directly to neglect of this 
"secondary" effect. But at any rate, it was generally 
conceded that with a suitable factor of safety the 
interactive effect would be negligible. Consequently, 
the usual approach taken was to consider the stress 
in the member to be made up of two parts; that due 
to bending and that due to the compressive load. 
This greatly simplified the task of estimating the 
dimensions of the cross section of the member. 

Wilbur and Orville Wright appear to have neglected 
the coupling effect during detail design of the spars 
used on their 1903 machine. This conclusion is 
supported by the calculation shown in Figure 12 and 
is dependent on use of the governing equation in the 

S = My/I + P/A 

where S = total stress in the spar in lb/in 2 
My/I = maximum bending stress 
P/A = compressive stress 
M = maximum bending moment in in-lb 
y = half beam thickness = h/2 where h 

denotes span thickness 
I = moment of inertia = bh 3 /12 where b 

denotes spar width 
P = compressive force in-lb 
A = area of spar cross section = bh 

For purposes of estimating the dimensions of the 
spar cross section it is convenient to simplify the 
expression to that shown at the right of the first line. 
Since the spar thickness, h, was set at 1.18 inches by 
the Wrights' choice of an airfoil and the spar location, 
the spar width, b, is the only unknown quantity in 
this expression. Thus, the spar width can be calculated 
directly by solving for b as shown on the second line. 
However, some explanation of how the values for 
bending moment, compressive force, and stress were 
selected so as to yield the indicated spar width may 
prove enlightening. 

While at Kitty Hawk on September 23, 1900 (Ref- 
erence 13), Wilbur wrote a letter to his father in which 
he states; "I am constructing my machine to sustain 
about five times my weight and am testing every 
piece." This often-quoted passage from the Wrights' 
correspondence has been interpreted by some writers 
to literally mean five times Wilbur's weight of 140 
pounds. Such an interpretation is meaningless in 
design of an aircraft. A more meaningful implication 
is that the Wrights planned to use a factor of safety 

Structural Design 


of five when designing their machines. However, 
what is now known as a load factor was often mistak- 
enly called a safety factor in the formative days of 
aviation. The different meanings of these two terms 
was to remain more than mildly controversial until 
finally resolved in the 1920s. 

The spars of the Wrights' 1903 machine were made 
of kiln dried spruce capable of sustaining a compres- 
sive stress of 6000 pounds per square inch. Reducing 
this value by a factor of 5 gives an allowable stress of 
1200 per square inch and this reduced value was used 
in performing the calculation shown in Figure 12. 

Wilbur and Orville Wright had an especially keen 
appreciation of material properties, as evidenced by 
their preference for West Virginia white spruce and 
second growth ash. Orville later wrote: "We figured 
we would get a slightly increased strength with the 
light spruce because with the same weight the cross 
section would be greater" (Reference 14). Such rea- 
soning is flawless when it comes to design of com- 
pressive members like spars or interplane struts that 
can fail by buckling. 

It may be recalled that the rear spar was estimated 
to carry 81 percent of the load when the center of 
pressure is at the 50 percent chord line. Calculations 
made using values of the bending moment and com- 
pressive force corresponding to this particular case 
yielded spar widths considerably smaller than the 
actual width of 1.68 inches. On the supposition that 
the Wrights' practical approach to problems may have 
prompted them to design the spar to carry the full 
load, the calculation was repeated using appropriate 
values for bending moment and compressive force. 
When used with the allowable stress of 1200 pounds 
per square inch mentioned earlier, this combination 
of values yielded the result shown in Figure 12 which 
agrees closely with the actual spar size used on the 
1903 machine. The indicated margin of 7.3 corre- 
sponds to the design load factor for the spar when 
subjected to the reduced loading of 81 percent. 

To further verify the meaning of Wilbur's statement 
to his father, it was decided to check the load factor 
used in sizing the rigging wire. This calculation is 
shown in Figure 13. In this case, the material was 
0.091 in diameter hard drawn steel wire listed as 
having a breaking strength of 1300 pounds in several 
early handbooks. As shown in the Maxwell's Diagram 
of Figure 9, the maximum applied load carried by the 
truss wires is about 240 pounds. The design margin 
may be determined by dividing the breaking load by 
the maximum applied load. The margin of 5.4, 
corresponding to the design load factor, is seen to be 
in close agreement with the previous conclusion of 5. 
The difference can be attributed to the Wrights' 
choice of the nearest standard wire size that satisfied 
their requirements for safety. The next smaller wire 


S = 

jyu + -L = 

bh : 


b = 

6M P 1 «Q in 

_ + = 1.69 in 

2 Sh 


b = 1.68 in 


13. Load factor — Rigging wires 

P = 180 lbs 

70 in 

P = 180 lbs 



Column Load P = — f ■ 

cr i ' 

Material Spruce 

E = 1.6 x 10 6 psi 
Let P cr = 5P - 900 lbs 

i - bh 3 P C r L 



.07 in 

12 I 

b = =- = 1.99 in 

h 3 

b = 1.94 in 

Margin = 4.9 

14. Determination of strut width 

size of .080 inches has a breaking strength of 1000 
pounds, which would have reduced the design load 
factor to 4.2. 

Wilbur and Orville Wrights' design treatment of 
the interplane struts on their 1903 machine provides 
a convincing example of their design ingenuity and 
familiarity with stress analysis. As shown in Figure 
14, the gap between the upper and lower wings is 70 
inches. Moreover, from Maxwell's Diagram the most 


Howards. Wolko 

highly loaded strut is seen to carry a compressive load 
of around 180 pounds. When such a long thin member 
is loaded in compression it can fail, without warning, 
by what is known as column buckling. Since the 
interplane struts on the 1903 machine function as long 
columns, their design is governed by an equation that 
can be written. 

TT 2 EI 

where P, r = Critical load at which failure is imminent 
E = Elastic modulus psi (For spruce E = 1.6 
million psi) 

bh 3 


I = Moment of inertia in 4 : 

b = Strut width in 

h = Strut thickness = .75 in 

L = Length of column in 

The Wright brothers clearly realized that a 70-inch- 
long column would require a substantial cross section 
in order to sustain a load of 180 pounds without failing. 
But a column of sufficient size would add weight and 
increase drag. Consequently, they sought an alternate 

Midway between the upper and lower wings of the 
1903 machine is a horizontal wire in the plane of the 
front spars. It passes through the center of all highly 
loaded struts and is wrapped and sized at the points 
of entry and exit. The function of this wire is to 
stabilize the struts against lateral deflection and thereby 
increase the allowable column load. In effect, the 
wire halves the strut length and increases the load 
factor to 4.9, a value consistent with the Wrights' 
design criterion. A second wire in the plane of the 
rear spars stabilizes the rear struts and results in a 
comparable advantage. The fact that none of the 
struts failed during the Wrights' proof-of-concept 
flights in the gusty winds of December 17, 1903, 
attests to the effectiveness of this design solution. 

Concluding Remarks 

This article has attempted to shed some light on how 
Wilbur and Orville Wright went about design of the 
Flyer. Wilbur, the self-appointed spokesman of the 
two, had planned to publish a description of their 
design procedure, but his untimely death from typhoid 
in 1912 left that task undone. Consequently, it has 
been necessary to reconstruct their approach to aircraft 
design. To add credence to this reconstruction only 
information extracted from The Papers, from engi- 
neering handbooks published prior to 1900, and from 
engineering drawings donated to the National Air and 
Space Museum by the Orville Wright Estate was 

used. The close agreement of all numerical results 
with evidence extracted from these sources suggests 
the Wrights used methods at least comparable to those 
presented here. It was particularly encouraging to 
note that none of the methods required a facility with 
mathematics beyond that normally acquired at the 
high school level. 

The Wrights' ability to prepare realistic design 
estimates reveals a fundamental appreciation for the 
way aerodynamic, propulsive, and structural consid- 
erations interact during aircraft design. Their under- 
standing of this interplay enabled them to set attain- 
able goals for power, thrust, and weight that served 
to guide them along an otherwise uncertain path. 
Moreover, their recognition that under idealized con- 
ditions flight loads could be approximated by consid- 
ering the external forces to be in equilibrium with 
the weight eliminated the need for resort to trial and 
error methods for sizing structural components. 

By today's standards the 1903 Flyer indeed was a 
most marginal machine, but it was not the product of 
chance or of tinkerers popular legend would have us 
believe. As this work indicates, the Flyer was a 
consequence of practical men endowed with extraor- 
dinary engineering perception. 


1. It should be noted in passing that the lift equation in 
current usage is 

L = |(Sc,V 2 ) 

where p = free air density in slugs/ft 3 

V = velocity in ft/sec 
At the sea level conditions of the Outer Banks, p is 
approximately 0.002377 slugs/ft 3 so that this expression 
may be written 

L = 0.001188 Sc,V 2 
The lift equation used by the Wrights was written 

L = 0.0033 Sc,V 2 with V = velocity in mph 
Converting the units of velocity to ft/sec gives the 
Wrights' equation in the form 

L = 0.0033(3600/5280) 2 Sc,V- = 0.001534Sc,V- 
which is in remarkably close agreement with current use. 
Doubting Thomases who believe the Wrights were 
tinkerers may find a moment's reflection on this 
agreement enlightening. 

Howard S. Wolko graduated from the University of Buffalo 
as a Mechanical Engineer and holds a D.Sc. in Theoretical 
and Applied Mechanics from George Washington Univer- 
sity. He served as a research engineer in experimental 
mechanics with Cornell Aeronautical Laboratory, a stress 
analyst and Head of Structures Research with Bell Aircraft 
Corp., Head of Solid Mechanics with the Air Force Office 
of Scientific Research, and was Chief of Structural Me- 
chanics and Reentry Structures with the Office of Advanced 
Research and Technology, NASA. He also was Professor 
of Mechanical Engineering at Texas A&M University, and 
was Professor and Chairman of the Mechanical Engineering 

Structural Design 


Department at Memphis State University. He currently is 
Special Advisor for Technology to the Aeronautics Depart- 
ment of the National Air and Space Museum. 

Dr. Wolko is a member of Pi Tau Sigma, Sigma Xi, the 
Society for Engineering Science, and the Society for Ex- 
perimental Stress Analysis. He is listed in Who's Who in the 
South and Southwest and American Men and Women of Science. 
He has published numerous technical papers. 


1. Wright, Wilbur, and Wright, Orville. The Papers of 
Wilbur and Orville Wright, Including the Chanute-W right 
Letters and Other Papers of Octave Chanute. Edited by 
Marvin W. McFarland. 2 vols. New York, 1953. Hereafter 
referred to simply as The Papers. 

2. The Papers, vol. 1, p. 102. 

3. Bisplinghoff, R. L. The Structural Engineering Practice of 
the Wright Brothers, in Aeronautica, Fall-Winter, 1956. 

4. The Papers, vol. 1, p. 54. 

5. The Papers, vol. 1, p. 44. 

6. The Papers, vol. 1, p. 290. 

7. The Papers, vol. 1, p. 313. 

8. The Papers, vol. 1, p. 386. 

9. The Papers, vol. 1, p. 135. 

10. Baker, M.P., The Wright Brothers as Aeronautical 
Engineers, Annual Report of the Smithsonian Institution, 
1950, p. 215. 

11. Judge, A. W., Design of Aeroplanes. London: Sir Issac 
Pitman and Sons, 1917, p. 156 

12. Ibid, p. 131. 

13. The Papers, vol. 1, p. 26. 

14. The Papers, vol. 1, p. 1106. 

1Q6 Howard S. Wolko 

ISBN D-fl?47M- t 17 c ]-q 



z x^iu^v m \ z ^ — -^ w ± c/> ± — </> 



> XiOmsJX 2 






g w KioiiniiisNrNViNOSHims^s3 1 y vy a h z li b rar i es w smithsonian z 








N N0linillSN| 2 NVIN0SHllWS C/> S3iyvyan 2 LIBRARlEs' /, SMITHS0NIAN2iNSTITUTI0N"'N0linillSNI_NVIN0SHillAIS 





S Xi£os£iX 



— t 


1 L 

*U>> : . . 


~ MBL. « 




z ^ ■ «? ^^ 5 ^_ ' « i 

< w 


2 <"" , 2 


2 ». 2 w • z -,. CO -, 



2 *\^v<&. -< <SM/!i>, 



i w N0iiniusNi_NviN0SHiiws^S3 1 avaa nf u brar i es^smithsonian ^institution ^oiiniiismf nvinoshiiws 


w N0linillSNI 2 NVIN0SHll^S M S3iaVaan z LIBRARIES^SMITHS0mAN-|NSTITUTl0N C/, N0linillSNrNVIN0SHllWS W ! 

W i ^^^ 1 ^*»r 5 v ? w*P7 ^ ^ Tsar = 


< v 2 x5?F^\ < \ £ -^ 5? : _ £ *v 5? * ^ ^ -. z 

s > -^^ i >\y- =■> - - - \$*^y %fr / t ^ z \^^/ g